AI Visibility Statistics From 1,352 Real AI Answers

Brett D. Henderson on September 22, 2026

AI Visibility Statistics From 1,352 Real AI Answers

These AI visibility statistics come from 1,352 real answers about 17 local businesses, and the headline number is 18.1%. That is how often the AI engines named the business we were asking about, which is all AI visibility means in practice: whether an assistant says your name when a customer asks it for a recommendation. We put one question to five engines, "What [service] company do you recommend in [city]?", across 79 cities between August 29 and September 22, 2026. In 162 of the 272 service and city combinations we tested, no engine named the business at all.

Key takeaways

AI named the business being measured in 18.1% of answers (245 of 1,352). The typical business in the group appeared in 17.1% of its own answers.
14 of the 17 businesses were named in fewer than half their answers. Two were never named once, by any engine, in any city.
In 162 of 272 service and city combinations, no engine named the business. That is 60% of everything we tested.
When AI did name a business, it named it first 51% of the time. Being mentioned at all is the hard part.
AI named about 3.5 businesses per answer, while BrightLocal's 2026 consumer research found 72% of buyers look at three or fewer before they decide.

How we measured this

Every number here traces back to one question, asked the same way every time: "What [service] company do you recommend in [city]?" We ran it for each service a business actually sells, in each city it actually serves.

Five engines answered: ChatGPT, Google AI Mode, Gemini, Claude and Perplexity. Grok appears in our report template but was never run, so every figure on this page reflects five engines, not six.

That worked out to 1,360 queries. We got 1,352 answers back. The eight that returned nothing were dropped rather than counted as failures, because an engine that times out has not made a judgment about anyone.

An engine that answered and named no businesses at all is a different thing, and those count as a miss. That distinction matters when you set our rates beside other published AI visibility data, so we are stating it up front.

One of the 17 reports used an earlier version of our format, covering a single service across eight markets with earlier model versions. We left it in and are telling you it is there.

If you want to see what the underlying websites looked like, we published audit findings from 41 local business sites separately. This piece measures what the AI said. That one measures what it had to work with.

GetLocalLeads.AI, an AI visibility agency for local service businesses, runs this same test every month.

Across all 17 businesses, AI named the business in 245 of 1,352 answers. That is 18.1%. The median business, the one sitting in the middle of the group, landed at 17.1%, so the average is not being dragged around by one outlier.

Underneath that, the spread is wide and mostly bad. Fourteen of the 17 were named in fewer than half of their answers. Ten of the 17 came in under a quarter. Two were never named at all, in any city, for any service, by any of the five engines. Eleven of the 17 had a competitor that AI named more often than them, and ten of the 17 reports came back with a Critical verdict.

It is fair to ask whether 18.1% is a bad number. Nobody expects to be named in every answer, and there are only so many slots. But flip it around for the businesses we measured.

For every ten customers who asked an assistant for a recommendation, roughly eight heard a list that did not include the business. Those eight never bounced off its website or ignored its ad. They were handed three or four names, none of them its own, so it was never in the running. There is no impression count anywhere that would show an owner that happening.

Here is the part that changes how you should read all of it. Fourteen of the 17 businesses were ranked number one by at least one engine, at least once. The knowledge is in there.

These are not businesses the internet has never heard of. They are businesses the engines know inconsistently, which is a different problem with a different fix. If you have wondered why you're missing from answers where you clearly belong, that gap between "known" and "named" is usually where it lives.

Seeing where you land is the first thing we do for a new client.

The same business, different city, different answer

Geography moved these numbers more than anything else we looked at.

We tested 96 business and city pairs. In 44 of them, 46%, the business was never named for any of its services in that city. Not ranked low. Absent.

The pattern inside that was consistent across the group. For businesses measured in more than one city, the strongest city averaged 45% of answers while every other city combined averaged 16%. Twelve out of twelve had a strongest city that beat the rest of their footprint. Eleven of 13 multi-city businesses had at least one city sitting at zero.

One Georgia business shows the shape of it clearly, and it is one case rather than a rule. Across its two home cities it was named in all 30 answers, every service, every engine, without a miss. In the largest expansion market it serves, it was named in none of the 15.

Businesses measured in a single city averaged 40.4% against 17.1% for multi-city businesses, but that comparison rests on only four businesses and 57 answers, so treat it as a direction, not a finding. The useful version is simpler: for the businesses we measured, AI visibility thinned out fast away from wherever the business is physically rooted, which is the same problem that makes cities without offices hard to win in ordinary local search.

We map this city by city before recommending anything.

Not every engine answers the same way

Gemini named the measured business in 28.8% of its answers. Google AI Mode managed 17.1%, Perplexity 16.2%, Claude 14.7%, and ChatGPT 14.0%. Google AI Mode is the conversational version of Google search, the one that answers in paragraphs instead of a page of blue links.

Before reading too much into that, here is the caveat that belongs in the same breath. Gemini names about 4.7 businesses per answer. ChatGPT names about 2.7. A longer list gives everybody more chances to appear, so some of Gemini's lead is generosity rather than knowledge.

Claude did something none of the others did at any scale: in 11.8% of its answers it declined to name a single business. An answer with no recommendation in it is still an answer a customer walked away from.

Each engine, taken alone, missed about six of the 17 businesses completely. Only seven of the 17 were named at least once by all five engines. Five of the 17 were named by one engine or none.

The practical read on all this is that engine choice is not a strategy decision you get to make. Your customer makes it, based on whatever app is already on their phone, and you have no visibility into which one they opened. Treating any single engine as the target means accepting that you are invisible to whoever picked differently.

The number that surprised us most: the average business had a 35 percentage point gap between its best engine and its worst. Same business, same question, wildly different answers depending on which assistant the customer happened to open. Anyone optimizing for one engine is reading a single column of a much wider table, which is why we keep saying beyond ChatGPT is where the work is.

Our audit covers all five engines, not just the one everyone names.

Being named is a switch, not a ladder

This is the finding we would keep if we had to throw out the rest.

Of the 245 times AI named one of these businesses, 124 of them, 51%, put it in the first position. Eighty-six percent landed in the top three. The average position across every mention was 2.0.

Now set that beside the misses. In 162 of 272 combinations, nobody was named at all. Only 10 of those 272 combinations had every engine name the business, which is 9% of the combinations where even one engine did.

Put those two facts together and the shape is clear. There is almost no middle. A business is either absent from an answer or sitting at the top of it. Very few land at four or five.

The engines do not even agree with each other on who belongs first. Across all 272 combinations, every engine picked the same number one business only nine times, 3.3%. We matched on exact business names, so spelling variants counted as disagreement and the true figure is a little higher. It is still tiny.

Traditional search trained everyone to think in ladders. You are on page two, you work to page one, you work into the top three. AI answers do not have a page two. There is no ranking below the answer that you are slowly climbing out of, because a customer never sees a fourth screen of ChatGPT results. Understanding how AI chooses a business matters more than any ranking report, because the whole game is getting inside the answer at all.

Why position four does not exist

Across 1,352 answers, the engines named 4,715 businesses. That works out to about 3.5 per answer.

BrightLocal's 2026 consumer research found that 72% of consumers look at three or fewer businesses before deciding, and only 1% look at more than ten.

Line those up. The assistant hands over three or four names. The buyer checks three or fewer. If you are the fourth name in an AI answer, you are already outside what most people will look at, and you are competing for a slot that shrinks every time the model writes a shorter list.

Of all 4,715 business names the engines produced across our tests, 245 belonged to the business we were measuring. That is 5.2%. The other 94.8% were competitors, and reports averaged 107 different companies named apiece.

One more thing worth knowing: 12 of the 49 service lines we tested were never named in a single city. A business can look fine on its headline rate while one of the services it sells is invisible everywhere it operates, which is something a single-number score will never surface. Getting cited has to happen service by service, not once for the whole company.

If you want the shortlist worked out for your market, book a call.

How this compares to the bigger studies

We are not the only ones measuring this, and the larger studies are worth knowing about. Most local SEO statistics you will find still measure Google: map pack appearances, review counts, click share. Very few measure what the assistants actually say, and the ones that do tend to study large chains.

SOCi's 2026 Local Visibility Index, published in March 2026, analyzed roughly 350,000 business locations across 3.2 million AI queries. It found brand locations appeared in AI recommendations 6.5% of the time, with Gemini recommending brands 11% of the time, against 36% for Google's local 3-pack. Coverage of that index also reported ChatGPT at 1.2% and Perplexity at 7.4%, though those per-platform splits were not on SOCi's own public page.

Our 18.1% is much higher than their 6.5%, and the reason matters. We asked about each business using its own service, in its own city. That is the most favorable prompt a business could get. SOCi sampled far more broadly. A targeted question is the ceiling, not the average, so read our figure as the best case and theirs as the wider picture.

The two datasets agree on something neither of us set out to prove. Gemini names the most businesses, ChatGPT the fewest. Different methods, different scales, same order.

The universe is different too. The average brand in their index had about 67 locations. The reader we built this for has one.

We publish our method so you can hold us to it.

What these numbers cannot tell you

Seventeen businesses is a small group. These are AI visibility statistics from the businesses we measured, and nothing here supports a claim about local businesses in general. Where we break the group into segments the samples get smaller still, down to four businesses in one comparison.

Every report is a baseline, captured once. None has been re-run, so nothing on this page measures improvement, or shows any work we did having an effect. For what a change over time actually looks like, our client's climb is a separate piece with its own evidence.

These businesses were also not chosen at random. They were measured for sales conversations and client work, which is its own kind of selection.

And engines vary between runs. Ask the same question tomorrow and some of these numbers move, so part of every rate here is noise rather than signal.

We publish it anyway, because a number you can check beats a rounder number you cannot. Ask us for the figures behind any claim we make.

Frequently asked questions

In our testing, AI named the business being asked about in 18.1% of 1,352 answers. SOCi's 2026 index of multi-location brands put the figure at 6.5%. Where a given business lands depends heavily on how the question is asked, and a targeted prompt about its own service in its own city is the best case.

Usually because it does not have enough consistent information to be confident naming you. In our data, 14 of 17 businesses were named first by at least one engine, so the information often exists but is not reaching every engine. You can check your business in a few minutes.

Which AI engine recommends local businesses most often?

Gemini, at 28.8% of its answers, followed by Google AI Mode at 17.1%. ChatGPT was lowest at 14.0%. Gemini also lists more businesses per answer, so some of that lead comes from longer lists rather than better knowledge.

Is being named second or third good enough?

Third is usually fine, fourth usually is not. AI names about 3.5 businesses per answer and most buyers look at three or fewer, so the bottom of a list gets very little attention.

How do I check whether AI recommends my business?

Ask each engine which company it recommends for your service in your city, then write down every business it names and where you appear. Run it for each service and each city separately, because both change the answer. That is exactly the method behind this page, and AI visibility work starts with running it properly.

Run it on your own business this week

The method on this page is not proprietary. Open all five engines, ask about your own service in your own city, and write down who gets named. Ten minutes will tell you which of these numbers is yours, and the result is usually more specific than anything a monthly report has told you. If you would rather see the whole picture across every service and city you serve, GetLocalLeads.AI does this for a living.

AI Visibility Audit: What 41 Local Business Sites Revealed

Brett D. Henderson on September 21, 2026

AI Visibility Audit: What 41 Local Business Sites Revealed

An AI visibility audit checks whether AI tools like ChatGPT, Gemini and Google's AI Overviews can read your website, believe what it says, and recommend you to a customer. In August and September 2026 we ran one on 41 local businesses across North America. Every one of them explained its services clearly. Not one scored above 41 out of 100, and every site had urgent fixes waiting. What held them back was small, visible and fixable.

Key Takeaways

The average score across our 41 audits was 34 out of 100, and the best score anyone reached was 41.
Access was mostly fine: 71% of the sites passed the crawl-access check. Trust is where they lost their points.
Two out of three of the businesses we audited wrote their own business name more than one way across their website and public profiles.
Five out of six of them had an error a customer could see on the page, from typos to leftover template text.

What does an AI visibility audit check?

An AI visibility audit is a review of the signals AI tools use to decide whether your business is real, what it does, and whether it is safe to recommend. Ours looks at six areas:

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found AI named the business being asked about in just 18.1% of answers, 245 out of 1,352. 2 of the 17 businesses were never named at all, by any engine, in any city.

Whether AI crawlers can reach your pages at all, through your robots.txt file (a short file that tells bots where they may go) and your sitemap (a list of your pages).
Schema, which is labeling in your page code that tells machines what each page is about.
Whether your name, phone number and other facts match everywhere they appear.
Proof a machine can verify, like reviews, named people and finished work.
How your content is organized.
Anything visible on the page that undercuts trust.

Each audit ends in a written AI visibility report and an AI visibility score from 0 to 100. That score measures something different from the free online checkers. Those tools count how often an AI mentions a brand. Ours measures whether your website gives AI a reason to mention you in the first place.

This matters more every month. BrightLocal's Local Consumer Review Survey 2026 found that the share of consumers using AI to find local businesses climbed from 6% in 2025 to 45%. If the idea is new to you, our guide to AI visibility covers the basics. The 41 businesses in this study were mostly plumbers, spread across at least 18 US states and one Canadian province.

If you want your own numbers, our free audit runs these same checks.

Every business explained its services. None scored above 41.

Every one of the 41 audits rated service clarity as strong, and the average score was still 34.2 out of 100. The median was 35. The lowest score was 24 and the highest was 41. These owners knew what they sold and said so plainly. A clear description of services was the one thing every site already had, which makes it the one thing none of these audits needed to fix. The trouble sat somewhere else.

The scores bunched low. Of the 41, 28 (68%) scored 35 or below, and 38 (93%) scored under 40. The 28 plumbing companies in the group averaged 34.9, a touch higher and still well short of halfway.

Every site also had urgent fixes, the problems we flag to correct before they cost a business trust. Across the 41 audits we logged 208 of them, an average of about five per site, and every site had at least three.

Here is the part that surprised us. Getting AI to the website was mostly handled. Of the 41 sites, 29 (71%) passed our crawl-access check, and 33 (80%) pointed bots to their sitemap correctly. Our piece on crawler access explains the two files involved if you want the mechanics.

Trust told a different story. Only 5 sites (12%) earned a strong rating for technical AI readability overall. Not one of the 41 earned a positive rating on AI trust signals: 38 were rated "needs work" and 3 were rated "bad." The door was open. What AI found inside gave it little reason to vouch for anyone.

That pattern repeats through every finding below. Access is mostly solved for these businesses. Belief is where they lose, and the sections that follow show exactly where it leaks.

The fastest way to see where your own site lands is a short call to look at it together.

How many businesses wrote their own name more than one way?

Of the 41 businesses, 28 (68%) wrote their business name more than one way across their website or public profiles. On their own websites alone, 22 (54%) did it. Among the plumbers, 17 of 28 (61%) had their name written two or more ways on their own site.

It rarely looked dramatic. An ampersand in the header and the word "and" in the footer. "Inc." on the About page and nothing on the homepage. A short nickname in the body copy. A parent company's name on one page and the trade name on the next.

Why does this matter to AI? An AI system keeps one record for each business, called an entity: a single thing with a name, a location and a set of services attached. When your site calls you three different things, it hands that system three candidates and no way to be sure which one is you. Our guides to name consistency and entity SEO go deeper on how that record gets built.

The same split happens across websites. Six of the businesses (15%) spread one brand over more than one domain, which divides the same signals between two addresses.

This check takes five minutes. Search your own site for every version of your name, then pick one. To keep it from drifting again, set up a single source of truth: one document holding your exact business details, which every page, profile and listing copies from.

A quick call is all it takes to see how your own site reads to AI.

How common were typos and leftover template text?

Thirty-four of the 41 sites (83%) had at least one visible copy error: a typo, placeholder text, stray characters or garbled symbols that any customer could see.

Spelling and grammar were the most common. In their own copy, 17 of the 41 (41%) had spelling or grammar errors. Counting errors inside the testimonials they chose to display, the number rises to 23 (56%).

Leftover template text was the most striking. Nine sites (22%) still had placeholder text live on the public site. A Florida plumber had a service page showing lorem ipsum, the fake Latin filler designers use in mockups, with "Author Title" printed where customer testimonials belong. A Kentucky plumber's page still carried the website builder's default line: "This is a paragraph where you can add any information you want to share with website visitors." A Florida restoration company had the words "New Paragraph" sitting on its homepage.

Some errors landed on the very words the business sells. A Missouri plumber's footer misspelled one of its own services and repeated the misspelling three times.

We did not measure whether errors like these change what AI recommends, so we will not claim they do. What we can say is simpler. The customer reads the same page the AI reads. A buyer judging website credibility sees placeholder text and a misspelled service and draws the obvious conclusion: nobody is looking after this site. That impression lands before the phone ever rings.

The check: read your footer and your full service list out loud, slowly. Typos hide in the places you stopped reading years ago.

If you would like a second set of eyes on your pages, get your audit.

What happens when a website contradicts itself?

Eleven of the 41 businesses (27%) published facts that contradicted each other, either on their own site or between their site and their Google listing. The conflicts covered review counts, phone numbers, license numbers, addresses and email addresses.

Review counts were the most common clash. Seven sites (17%) showed conflicting review numbers, and all seven were plumbers, which works out to 25% of the plumbing group.

Each example sounds small on its own. A Florida plumber showed two different review counts on the same site and listed three different license numbers. A California plumbing company's homepage claimed a few hundred Google reviews in one spot and nearly two thousand in another. An Oklahoma plumber called itself BBB (the Better Business Bureau) accredited while its BBB listing said otherwise.

One conflict costs calls directly. A Colorado company displayed one phone number, but its "Call Today!" button dialed a different one. Every customer who tapped that button reached a number the page never showed them.

A system that finds two answers to the same question cannot tell which one is true, so it has good reason to trust neither. People react the same way. Two review counts on one page read like a number nobody checked.

The check: open your site on your phone and tap every phone link. Then compare every review count and license number on the site with each other and with your Google listing.

When you are ready to see what your own site shows, book a call.

Why does so much proof look borrowed?

Proof is what separates a business that says it does good work from one that shows it. On these sites, much of that proof looked recycled or was missing entirely.

Start with testimonials. Of the 41, 13 (32%) repeated the same testimonials on one page or across several pages, and among plumbers the rate was 36%. A buyer who meets the same three quotes on the homepage, the service page and the contact page stops believing any of them. Fresh, specific real reviews carry more weight than a polished trio on repeat.

Finished work was missing almost everywhere. Of the 41, 35 (85%) had no case-study, results or project page at all. For a picture of what that kind of proof can look like once it is written down, see our local SEO case study.

People were missing too: 27 businesses (66%) showed no leadership or team bios. Person schema, the code label that tells machines who is behind a business, was present on only 3 sites (7%). An AI system deciding who to trust had no one to point to.

Claims ran ahead of evidence. Nineteen sites (46%) stated a guarantee or warranty, and only 7 (17%) said what it actually covers. Ten (24%) made unqualified claims such as "#1," "the best" or "100% guaranteed" with nothing on the page backing them, and 5 of those (12%) called themselves #1 or the best outright.

Outside proof was hard to confirm. For 26 of the 41 (63%), the audit could not verify their review profiles from the website itself. Those reviews may well exist. The site simply gave neither a machine nor a buyer a clear path to find them.

We run this audit free, and it checks every item above for your own site.

Which answers does AI look for that these sites left out?

Buyers ask the same handful of questions before they pick up the phone, and most of these sites gave those questions no home.

That gap matters because of how an assistant answers. When a customer asks it something, it looks for a page that already answers the question in plain words. A clear question as a heading, with a two-sentence answer under it, hands the assistant a ready quote. A site without one leaves it to quote somebody else.

FAQ pages were rare. Only 8 of the 41 (20%) had a dedicated FAQ page, and 22 (54%) had no FAQ content anywhere. Of the 19 that answered questions somewhere, only 4 marked those answers with FAQ schema so machines could read them as questions and answers. Clear question-and-answer content is one of the most direct ways to get cited by an assistant.

Price was a blank. Of the 41, 36 (88%) published no pricing at all. When a customer asks an assistant what a job costs, these sites gave it nothing of their own to quote.

The labeling was thin. Organization schema, the label that tells machines who a business is, was present on 20 of the 41 (49%) and partial on 3 more. For 12 sites (29%), the audit confirmed none of the six core schema types it checks. Our plain-English guide to schema markup shows what those labels look like.

Book a call and we will walk through which questions your site already answers.

What should a local business fix first?

The good news in this data is that the most common problems are also the cheapest to fix. In order:

Create a single source of truth for your business details, and have every page, profile and listing copy from it.
Proofread your footer, service list and homepage, and delete any template text.
Make every phone number, review count and license number match, on the site and on Google.
Swap repeated testimonials for distinct ones, and put a named person behind the business.
Answer your top buyer questions on the page, then add schema so machines can read them.

Treat that first item as a style guide for your business identity: your exact name, address, phone, hours, services and logo in one place, copied from there every time.

Most of that list is an afternoon of work, and all of it belongs ahead of any new content spend. Once it is done, you can check whether AI recommends you and see what changed.

Get your audit to see this list ranked for your own site.

How we ran these audits, and what the scan cannot see

These findings come from 41 AI Visibility Audits that GetLocalLeads.AI, an AI visibility and digital marketing agency for local and multi-location service brands, prepared in August and September 2026: 39 in August and 2 in September. Each one reviewed a business's public website. These were businesses we audited, not clients we served, so nothing here describes results we produced.

Twenty-eight of the 41 were plumbing companies, some also offering HVAC. The rest covered pest control, roofing, landscaping, water treatment, restoration, appliance retail and several professional and personal services. Forty were in the United States and one was in Canada. Every percentage in this piece is a share of those 41 audits, not a claim about all local businesses.

The scan has limits, and we would rather name them. It reads public pages, so it cannot always confirm a Google Business Profile or a third-party review count. A few checks were not reported for every site, so we left them out. Duplicate navigation text that may come from the scanning process itself was excluded from every count.

Frequently asked questions

What does an AI visibility audit check?

An AI visibility audit checks whether AI tools can reach your website, understand what it is about and trust what it says. Ours covers crawler access, schema, name and fact consistency, verifiable proof, content organization and visible trust problems. It ends with a score out of 100 and a ranked list of fixes.

Do I need an AI visibility audit if I already rank on Google?

Ranking and being recommended are separate questions, and our audits measure the second one: whether your site gives AI a reason to trust it. That question matters more each year. BrightLocal's 2026 survey found 45% of consumers now use AI to find local businesses, up from 6% a year earlier.

What is a good AI visibility score?

In our 41 audits, no business scored above 41 out of 100, and the average was 34. A low score is normal and useful, because it arrives with a ranked list of what to fix. Free online checkers that show an AI visibility score usually measure something different: how often a brand gets mentioned.

How much does an AI visibility audit cost?

GetLocalLeads.AI runs a Free AI Visibility Audit. To qualify, you need a website, a name, an email and a company URL. The comprehensive audit carries a stated value of $997, and you can request yours by booking a call.

Can I check my own website for these problems?

Yes. The five fixes above double as a self-check you can run in an afternoon. To test whether assistants actually name you, follow our guide to checking whether AI recommends your business.

The pattern across these 41 businesses is oddly hopeful. The problems that cost them the most trust were the ones that cost the least to fix: a name, a typo, a phone link, a repeated quote. The fix starts with someone reading the site the way a stranger would, and our AI visibility services pick up from there. If you want that first read done for you, book a call and we will run your free audit.

More GetLocalLeads.AI case studies

Is SEO Dead? What AI Answers Really Changed

Brett D. Henderson on September 18, 2026

Is SEO Dead? What AI Answers Really Changed

Is SEO dead? No. But one specific version of it is finished, and pretending otherwise is a
large part of why business owners stopped believing marketing agencies. Content written to
collect clicks from question searches has lost most of its value. Work that gets your
business named when somebody is ready to hire has not. Those two things are moving in
opposite directions, and most articles on this question average them together.

Key Takeaways

The Pew Research Center found Google users clicked a search result in 8% of visits when an AI summary appeared, compared with 15% of visits when no summary appeared. The drop is real and it is measured.
That click loss sits almost entirely in question-shaped searches. Searches where someone is trying to hire a business still resolve to a list of business names.
Google's own documentation states there are no additional requirements to appear in AI Overviews or AI Mode. Anyone selling you a separate "AI search" product is selling you the same work twice.
What died is thin informational content built to harvest traffic, and the ranking report that never explained why the phone was quiet.

What is the strongest case that SEO is dead?

The case is better than most agencies admit, and it comes down to one thing: Google now
answers the question on the results page, so the click that was the entire point of ranking
never happens. Here it is without the softening.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found every engine picked the same number one business in only 9 of 272 combinations, 3.3%.

For twenty years the deal was simple. Somebody typed a question, Google showed ten links,
and one of those links was yours. Now Google writes the answer at the top of the page, and
the question gets resolved before anyone scrolls.

This is not a feeling. The Pew Research Center tracked the browsing of 900 U.S. adults
through March 2025 and matched it against search results collected that April. When an AI
summary appeared, people clicked a traditional search result in 8% of visits. When no summary
appeared, they clicked in 15% of visits, close to twice as often. Sessions that ended without
any click at all rose from 16% to 26%.

Then comes the number that should bother anyone selling search work. People clicked a link
inside the AI summary itself in 1% of visits. One percent. The standard reassurance from the
industry right now is "get cited in the AI answer and you will still get the traffic," and
Pew's data
does not support that sentence.

For one type of business, the conclusion follows cleanly. If what you sell is the page view,
if your revenue is advertising against traffic, this is an extinction event and you are right
to be furious. Publishers, affiliate sites, and content farms are saying SEO is dead because
for their model it substantially is. They are not being dramatic. They are describing their
own income statement.

That is the strongest version of the argument. Now here is where it stops applying to you.
GetLocalLeads.AI walks owners through this on a call most weeks.

What the click data is actually measuring

Read the Pew study carefully and you notice what it measured: what people did on Google
results pages, with and without an AI summary. It did not measure whether anyone booked a
plumber. It measured clicks on a page, which is a different thing from customers.

That distinction matters because AI summaries do not show up evenly. They show up where
someone asked a question. Local search firm Whitespark ran 540 queries across Houston,
Phoenix and Denver and found AI Overviews on 92% of informational-intent queries but only
15% of local-intent ones, while the local map pack, the block of business listings Google
shows under a map, appeared on 93% of those same local-intent searches. That work is from
May 2025, so treat the exact percentages as stale by now. The shape of the finding is the
durable part: the two features move in opposite directions depending on what the searcher
wants.

Say it as a rule instead of a statistic. The more a search looks like a question, the more
likely an AI answer consumes the click. The more it looks like somebody trying to hire
today, the more it still ends in a list of business names.

Now put that next to your own numbers. Informational traffic is what a publisher lives on.
For a roofing company, it was always the worst traffic on the site. Somebody reading "how
long does a roof last" is doing homework, not buying. Losing that visitor feels like losing
everything only if it was everything you had. If you are a service business, it was never
where your money came from. Worth a conversation if your phone has been quieter than your
rankings suggest it should be.

Will AI replace SEO?

No, and the reason is mechanical rather than reassuring. An assistant has no independent
knowledge of which plumber in your town does good work. It reads the same public web, the
same business profiles, and the same reviews that a search engine reads. Take the source
material away and the assistant has nothing to say.

Google states this outright in its
AI features documentation:
"There are no additional requirements to appear in AI Overviews or AI Mode, nor other special
optimizations necessary." The same page adds, "You don't need to create new machine readable
files, AI text files, or markup to appear in these features," and concludes that "the best
practices for SEO remain relevant for AI features in Google Search." If someone is quoting
you a separate line item for AI search on top of your existing search work, that quote is
worth a hard question.

Google also argues the traffic damage is smaller than it looks. Liz Reid, the VP who runs
Search, wrote in August 2025 that
total organic click volume
"has been relatively stable year-over-year." Hold that at arm's length. It is Google
describing its own product, using aggregate figures nobody outside Google can check, while
Pew measured actual sessions from real people. Two credible sources, two methods, two
answers. The honest read is that the input to search barely changed. What changed is where
the output appears, which is what
answer engine optimization
is about. See how we approach it before committing to anything.

Is local SEO dead?

No, and the answer here is flatter than for search in general. A local search still ends in
a short list of named businesses, and something still has to choose those names.

Whether that list is Google's map pack or three companies mentioned inside a chat reply, the
selection runs on the same inputs. The engine needs an accurate and consistent picture of who
you are: the same business name, address, phone number and service list everywhere it appears
online, a Google Business Profile that matches your website, reviews from real customers, and
a site a crawler can actually read. Get those wrong and you are not penalized so much as
skipped, because the system cannot describe you with confidence.

The stakes went up rather than down. BrightLocal found in 2026 that consumer use of AI tools
to find local businesses rose from 6% to 45% in a year. A search engine gave your customer
ten options and a map. An assistant gives them two or three names. Being seventh used to
mean a slow month. Being fourth in an AI answer means you were not mentioned.

The practical difference shows up in two searches a contractor sees every day. "How much does
a new furnace cost" is a homework question, and an AI summary will likely handle it. "Furnace
replacement Sarasota" is a hiring question, and it still produces a map pack or a shortlist.
One of those two searches was ever going to become a job. Our guide on
ChatGPT recommendations
covers how assistants build that shortlist. There is a free AI visibility audit if you want
the answer for your own business.

What actually died

Something did die, and naming it precisely is more useful than arguing about whether search
has a pulse.

In one of our internal studies of 41 local service businesses, 71% let AI crawlers in, but only 12% rated strong on technical AI readability and none earned a positive rating on trust.

Thin informational content died first. If a page existed to rank for a common question and
funnel a stranger toward a contact form, the AI answer now does that job for free and better.
Publishing forty articles a month about roof maintenance to catch stray searches is a
strategy with no future.

Ranking as the deliverable died next. Position was always a proxy for attention, and it is a
poor one now, because a page can sit at position three underneath an answer box that already
handled the query. A report showing you moved from eleven to six explains nothing about why
the phone did or did not ring. An HVAC company can hold the top organic spot for "how often
should I change my air filter" and take zero calls from it, because the searcher got what
they wanted without leaving Google. The ranking is real. The value attached to it is not.

Keyword counting died a long time ago and some people are still billing for it. Repeating a
phrase a certain number of times stopped working years back. Search systems read meaning now,
and stuffing gets you filtered rather than ranked.

The monthly report full of impressions died too, or it deserves to. Impressions, reach and
clicks were never the product. Calls, booked jobs and revenue were. Book a call if you want a
second opinion on what you are currently paying for.

When SEO is genuinely the wrong spend

There are situations where search work is the wrong purchase this quarter, and saying so is
more honest than a universal pitch.

The first is capacity. If your crews are booked six weeks out, or calls are already going to
voicemail during the day, more visibility pours water into a bucket with a hole in it. Fix
the answering problem or the scheduling problem first. The return is faster and it costs
less.

The second is foundations. A brand-new business with no reviews, an unclaimed Google Business
Profile and a site that blocks crawlers is not ready for content. Publishing articles on top
of a broken business identity wastes months, because the system still cannot tell who you
are.

The third is timing. Search work compounds over months. If you need leads next Tuesday, paid
search buys attention today and organic work does not. One tool fits one problem.

The rule underneath all three: search is the wrong spend when your real constraint is
capacity, foundations, or the calendar.

Does SEO still work for a small business?

Yes, with a condition attached rather than a promise. Search work pays off when the query you
are trying to win is one where a buyer picks a name. It does not pay off when you are chasing
question traffic that an AI answer now handles.

So change the question you ask your marketing. Not "where do we rank," which was always a
proxy. Ask how many people called after finding you, and where those people came from. That
number is harder to produce and it is the only one that pays anybody's payroll.

The work itself looks different than it did in 2019, though less different than the headlines
suggest. Make the site readable to crawlers and assistants. Make your business identity
unambiguous everywhere it shows up. Answer real customer questions directly on the page
instead of circling them for 800 words.

Part of that is what we call entity hardening, which means setting clear boundaries around
what your business is and shows online, so an AI tool forms an accurate picture of what you
do and, just as importantly, what you do not do. A roofer who also does gutters should be
findable for both and mistaken for neither. That is the core of how GetLocalLeads.AI, an AI
visibility and digital marketing agency for local service businesses, approaches
SEO, GEO and AEO together rather than as separate
products. The free AI Visibility Audit is a reasonable place to start.

Frequently Asked Questions

Is SEO dead in 2026?

No. Search engines and AI assistants both still have to choose which businesses to name, and
both make that choice from the public web. What has lost most of its value is content written
to collect clicks from question searches, because an AI answer now handles those questions on
the results page.

Will AI replace SEO?

No. An assistant has no private database of good contractors. It reads your website, your
business profile, your reviews and the directories you appear in. Google's own documentation
says the best practices for SEO remain relevant for its AI features.

Is local SEO dead?

Local search is the healthiest part of search right now. Hiring-intent queries still return a
map pack or a short list of business names, and AI assistants build their shortlists from the
same profiles and reviews. The main change is that the list got shorter.

Is SEO worth it for a small business?

It is worth it when you have the capacity to take the calls and the foundations to be found:
a claimed business profile, consistent information, real reviews, and a crawlable site. If
those are missing, fix them first. A
local SEO audit will tell you which situation
you are in, and we are happy to look at it with you.

How long does search work take to show results?

Traditional search results take months, which is why contracts exist. AI visibility moves
faster: movement is possible within 24 hours of a website update such as a corrected
robots.txt file (the file that tells crawlers what they may read), though results are not
guaranteed. Treat any fixed-timeline promise as a warning sign.

What to check before you cancel anything

Open ChatGPT or Google's AI Mode and ask for the best version of your trade in your town. Ask
three different ways. If your name never comes up, you have your answer, and it is not that
search is dead.

People have asked whether SEO is dead every year since at least 2016, and the keyword tools
still show dated versions of it going back that far. It returns because the answer keeps
changing shape, while the businesses that get named stay the ones somebody can describe
clearly. GetLocalLeads.AI can walk you through where yours stands.

ChatGPT SEO: How to Rank on ChatGPT

Brett D. Henderson on September 18, 2026

ChatGPT SEO: How to Rank on ChatGPT

ChatGPT SEO is the work of making your website readable, quotable, and consistent enough that ChatGPT names your business when a customer asks it who to call. It starts with one file on your own server: your robots.txt has to allow OpenAI's search crawler, or you are shut out of ChatGPT's search answers no matter how good your site is. After that, three things carry the weight. Answer real questions plainly on your pages, make every public mention of your business say the same thing, and test the result with the prompts your customers actually type.

Key Takeaways

OpenAI runs separate crawlers with separate controls. Blocking AI training does not block ChatGPT search, and a blanket block that catches the search crawler removes you from ChatGPT's answers entirely.
There is no submit button and no paid placement. You cannot apply to be recommended and you cannot buy the spot.
Assistants lift passages, not pages. A paragraph that only makes sense after the three paragraphs above it does not get quoted.
One location wins local prompts, not national ones. "Best plumber in Tulsa" is winnable. "Best plumber in America" is not, and it was never worth anything anyway.

Why Your Customers Are Asking ChatGPT Before They Call You

The phone is slower than the work deserves and nothing obvious changed. Same crew, same trucks, same five-star reviews, same ad spend. That is usually where an owner starts looking for a villain, and usually the villain is not there.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found when AI did name a business it named it first 51% of the time, 124 of 245 mentions.

What changed is where the question gets asked. BrightLocal's Local Consumer Review Survey, published in February 2026, found that use of generative AI tools for local recommendations rose from 6% the year before to 45%. In one year, asking an assistant went from a curiosity to something close to half of local searchers.

The shape of the answer changed with it. A results page gives a customer ten links and lets them pick. An assistant gives two or three names and a sentence about each. Position eleven on Google was always invisible, and everybody in the trades knew it. Not being in a three-name shortlist is the same kind of invisible wearing a new coat, except almost nobody has checked whether it is happening to them.

This is the problem GetLocalLeads.AI, an AI visibility and digital marketing agency for local service businesses, works on every day.

What ChatGPT SEO Actually Means

ChatGPT SEO is the practice of making a business's website and public information readable and quotable enough that ChatGPT surfaces and cites it when someone asks a relevant question. It sits inside the broader category of answer engine optimization, which covers every assistant, not only this one.

There is no form to fill in. OpenAI's own announcement of ChatGPT search says "Any website or publisher can choose to appear in ChatGPT search," and choosing to appear means setting crawler permissions on your own site, not applying to anybody. Nobody sells the spot either. If a vendor offers to place you in ChatGPT's recommendations, they are describing a product that does not exist.

Here is where most ChatGPT optimization advice goes wrong. Most guides on this topic tell you to go improve your Bing rankings, stated as settled fact. OpenAI describes ChatGPT search as drawing on "third-party search providers, as well as content provided directly by our partners," and OpenAI also runs its own search crawler. So the Bing instruction is an inference the industry made, not something OpenAI's own description of the product says. It might help. It is not the lever.

If you want the deeper version of how an assistant decides which name to say, we covered the signals in how ChatGPT recommends businesses. This piece is about what you do on Tuesday afternoon.

Worth a look at how the SEO, GEO and AEO work runs if you would rather hand this off.

Step One: Make Sure ChatGPT Is Allowed to Read Your Site

This is first because it is free, it takes ten minutes, and it is broken more often than anyone expects.

OpenAI publishes documentation for four separate crawlers, a crawler being a program that visits and reads your pages. OpenAI's crawler documentation names them: OAI-SearchBot/1.4 handles search, and OpenAI says it "is used to surface websites in search results in ChatGPT's search features." GPTBot/1.4 crawls content that may be used to train the underlying models. ChatGPT-User/1.0 visits a page when someone in a chat triggers it. OAI-AdsBot/1.0 checks the safety of pages submitted as ads.

The part nobody explains is that these are four different switches. In OpenAI's words: "Each setting is independent of the others – for example, a webmaster can allow OAI-SearchBot in order to appear in search results while disallowing GPTBot." A lot of small business sites got a plugin update two years ago with a friendly "block AI bots" toggle, and the owner flipped it because training an AI on their site felt like being robbed. Reasonable instinct. The toggle often writes a blanket rule that catches the search crawler too.

That matters because of the consequence OpenAI states plainly: "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links." Opted out means absent from the answer.

To check, type your domain followed by /robots.txt into a browser. You are looking for any Disallow line sitting under OAI-SearchBot, or a blanket disallow near the top that applies to everything. The three causes worth suspecting, in order: a security or privacy plugin with an AI-blocking setting, a leftover rule from when the site was on a staging server and disallowed the whole internet, and a developer handover that nobody audited afterward.

One nuance so you do not draw the wrong conclusion from your own testing. On the user-triggered crawler, OpenAI notes that "Because these actions are initiated by a user, robots.txt rules may not apply." That means ChatGPT can fetch your page live in a conversation while your site is still missing from its search answers. Watching it read your page once is not proof the door is open. For the file mechanics themselves, we wrote up robots.txt and llms.txt separately.

If that file is unfamiliar territory, this is a reasonable thing to have someone else look at.

Step Two: Write Pages a Machine Can Quote

Assistants lift passages, not whole pages. That one fact should change how your service pages are written.

Answer the question in the first two sentences under every heading. If a paragraph only makes sense after the three paragraphs above it, it cannot be pulled out and used, so it will not be. Say the answer, then explain it, then give the context. That is backwards from how most service pages read and it is the single cheapest structural fix available.

Use the customer's question as the heading. "How much does a water heater replacement cost in Tulsa?" is a heading a machine can match to a real query. "Our Water Heater Services" is a heading that matches nothing anyone has ever typed.

Give every real service its own page, named the way a customer says it. A single Services page listing nine trades gives an assistant nothing specific enough to attach your name to. When the question is about emergency sewer line repair, the page that gets cited is the one about emergency sewer line repair.

A concrete version of all this: a roofing page opening with "Founded in 2004, our family-owned team has proudly served the community" contains nothing an assistant can quote. The same page opening with "Most residential roof replacements in Tulsa take four to seven days, weather permitting" hands it a usable sentence.

Then put the proof next to the claim. License number, years in business, the towns you actually cover, your response window, in plain text on the page rather than baked into an image a crawler cannot read. Add schema while you are there, which is structured data, meaning labels in your page code that tell a machine which piece of text is the phone number, which is the address, and which is the hours.

Happy to walk through what this looks like on your own site.

Step Three: Make Every Mention of You Say the Same Thing

An assistant builds its answer from several sources at once. When your name, address, phone number, and list of services read differently across those sources, you stop looking like one confident business and start looking like three half-confirmed ones. The assistant does what a cautious person would do and names someone it is surer about.

In an internal study of 41 local service businesses, we found that 68% wrote their business name more than one way across their website and public profiles, so this is the step where most of the businesses we audited fell short.

So sweep the obvious places: your website, your Google Business Profile, the directories your trade actually lives in, and the old listings carrying a phone number you stopped using in 2019. NAP consistency, meaning your name, address and phone number matching everywhere they appear, and local citations, the directory listings that repeat those details, are each a project of their own, so this is the summary version. One name, one address, one phone number, one description, everywhere.

The usual objection is that a listing service already handles this. Sometimes it does. What it rarely handles is the listing you created yourself in 2016 under a slightly different business name, on a directory you have long since forgotten, still sitting there quietly contradicting everything else.

The underused half of this is stating what you do not do. GetLocalLeads.AI calls it entity hardening: establishing clear boundaries for what a business is and shows online, so AI tools can tell where you deliver excellent service and where you do not offer service at all. If nothing on the internet makes clear that you are residential only, an assistant that is unsure will not risk naming you for a commercial job. Saying no out loud makes the yes more credible.

That boundary work is the core of GetLocalLeads.AI's AI visibility and brand management service.

Step Four: Test It With the Prompts Your Customers Actually Use

Stop guessing and go ask. Working out how to get your business on ChatGPT starts with finding out whether it already names you. Open ChatGPT and run five prompts, substituting your trade and your city:

"Best [trade] in [city]"
"Who should I call for a burst pipe in [city] tonight"
"[Trade] near me open now"
"Is [your business name] any good"
"Who does [your most profitable service] in [city]"

Write down three things for each: whether you were named, who was named instead, and which pages the assistant cited underneath the answer. The citations are the useful part. They tell you which sources the assistant trusts in your market, and that list is usually shorter and stranger than anyone expects.

Two caveats so you read the results honestly. Start a fresh chat, and a logged-out one where you can, because your own history pulls the answer toward things you have already talked about. And run this monthly rather than daily. Answers move around on their own, and chasing day-to-day movement will make you crazy without teaching you anything.

Read the outcome in three buckets. Named and described accurately is a win. Named but described wrong is a consistency problem, and it is the easiest of the three to fix. Not named at all, with competitors cited from sources you have never heard of, tells you exactly where the work is.

If you would rather see the full picture than five prompts, the free, comprehensive AI Visibility Audit covers it: Get My Audit!

How Long This Takes, and What It Will Not Do

SEO for ChatGPT does not run on the same clock as Google SEO. AI visibility moves faster than traditional search optimization, and movement is possible within 24 hours of a website change such as an updated robots.txt file, though results are not guaranteed. An access fix can land quickly because it removes a block rather than building an argument. Clarity and consistency work behaves the way SEO always has, which is to say slower.

Now the ceiling. A single-location service business is not going to be named for "best HVAC company in America," and no amount of work will change that. The ground worth fighting for is the city-and-service prompt, because that is what a person with a broken furnace actually types. Winning that is smaller and worth considerably more.

Assistant answers are also probabilistic. The same prompt can return different names on different days and from different accounts, because the model is generating an answer rather than reading a fixed list. Anyone promising you a locked position inside an assistant's response is describing something that does not exist. What you can move is the odds, and the odds move a lot.

One more thing, said plainly because it saves people money. If your Google Business Profile is unclaimed or your website will not load on a phone, fix that first. ChatGPT is reading the same public information those channels feed, and there is no point optimizing the summary of a story nobody has written down yet.

Frequently Asked Questions

Does ChatGPT use Bing's index?

OpenAI's announcement of ChatGPT search does not name one. The company describes ChatGPT search as drawing on "third-party search providers, as well as content provided directly by our partners," and it runs its own search crawler on top of that. The popular "it is just Bing" shorthand is an industry assumption, not documentation. Allow OAI-SearchBot, stay indexable everywhere, and skip optimizing for one engine whose role you cannot verify.

No. There is no paid placement in ChatGPT's recommendations and no form to submit your business for consideration. OpenAI does document a separate crawler that checks the safety of pages submitted as ads, which is worth knowing so you do not confuse advertising with being recommended. Different systems, and the recommendation is not for sale.

How long does ChatGPT SEO take to work?

It depends what you fixed. AI visibility moves faster than traditional SEO, and movement is possible within 24 hours of a website change such as an updated robots.txt file, though results are not guaranteed. Removing a block is fast. Earning a mention where your name barely appears takes months of consistency work. Book a call for a straight read on which one you are dealing with.

Will blocking AI training hurt my ChatGPT visibility?

Not by itself. OpenAI states that "Each setting is independent of the others – for example, a webmaster can allow OAI-SearchBot in order to appear in search results while disallowing GPTBot." So you can refuse training and still appear in search answers. The risk is the blanket block many plugins write, catching the search crawler along with the training one. Open your robots.txt and read what it actually wrote.

Is ChatGPT SEO the same as ranking on Google?

Overlapping, not identical. A fast, crawlable site with clear content helps both, so the foundations transfer. The output differs: Google hands back a page of links to scan, and an assistant hands back two or three names. Google's own AI answers are a separate topic with their own rules.

One Last Thing

The cheapest item on this list is the one almost nobody checks. Before you commission another page, buy another directory listing, or sit through another agency pitch, open your own robots.txt file and read it. Some owners find a blanket block a plugin wrote for them two years ago, and undoing it costs nothing. Start there, then book a call when you want the rest of it handled.

AI Search Optimization: What Changes and What Does Not

Brett D. Henderson on September 18, 2026

AI Search Optimization: What Changes and What Does Not

AI search optimization is the work of making sure AI tools can find your business, read it correctly, and name it when a customer asks them a question. Most of that work is the SEO you are already paying for. Google says its AI answers run on the same ranking systems that produce the ordinary results, so the fundamentals were not thrown out. What changed is the shape of the answer. A list of ten blue links became a shortlist of two or three names, and that single change moves the goal from being clicked to being quoted.

Key Takeaways

Google states its AI features are "rooted in our core Search ranking and quality systems," which makes good SEO the entry fee for AI answers rather than a legacy expense.
The genuinely new part is the shape of the result. Ten links became two or three names, so the job moved from ranking on a page to being named in an answer.
For a local business, the AI answer is assembled from your Google Business Profile, your reviews, and your listings as much as from your website.
A page that is blocked from showing a search snippet cannot appear in Google's AI answers at all, which is the most common silent failure we see.

What is AI search optimization?

AI search optimization is the practice of organizing a business's website and public information so AI systems can read it, trust it, and repeat it accurately when someone asks them a question.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found Gemini named the business in 28.8% of its answers while ChatGPT managed 14.0%, though Gemini also lists more businesses per answer.

You have probably been handed four words for this one idea, which is most of why it feels like a new expense. AEO means answer engine optimization, or getting quoted in a direct answer. GEO means generative engine optimization, or getting cited by tools that write an answer from scratch. AI SEO is the loose catch-all. Search everywhere optimization is the newest label, and it just means people now start searches in places that are not Google.

Google settled the question in its own documentation: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Most of the ai search optimization vs SEO debate is a vocabulary problem wearing a strategy costume. Wikipedia's article on the subject records a Forrester analyst's 2025 assessment that these terms are "significantly, but not fundamentally, different from SEO," which matches what we see in client accounts. If you want the fuller comparison, we wrote one on answer engine optimization.

If the vocabulary is the part giving you a headache, a short conversation clears it up faster than another article will.

What actually changed?

People used to get ten links and pick one. Now a lot of them get a written answer with a few names in it, and never scroll.

The Pew Research Center measured this on Google itself. In a study published in July 2025, drawn from 900 U.S. adults and 68,879 searches during March 2025, users who saw an AI summary clicked a traditional search result on 8% of visits, against 15% of visits where no summary appeared. A link inside the summary got clicked on 1% of visits.

Google describes the same page differently. In its documentation on AI features, it says that "when people click from search results pages with AI Overviews, these clicks are higher quality (meaning, users are more likely to spend more time on the site)." Both statements can be true at once, and the practical reading is fewer visits that are further along in deciding.

The result got shorter, so being on the list stopped being enough and being named started to matter. Seeing where your business currently lands in those answers takes one call.

Which parts of your SEO still count?

Google was asked directly whether SEO is still relevant for generative AI search, and answered: "In short, yes! The best practices for SEO continue to be relevant because our generative AI features on Google Search are rooted in our core Search ranking and quality systems." That sentence is worth reading twice if you have been told your investment is dead.

In an internal study of 41 local service businesses, 71% let AI crawlers in and 80% declared a sitemap correctly, yet only 12% rated strong on technical AI readability.

Three groups of work, and each one gets a different verdict.

Some of it counts exactly as much as it did before. Useful content written for customers instead of for search engines. A site that loads on a phone in a truck cab. Clean crawling and indexing. Clear page titles and honest headings. Accurate information about what you do and where you do it. None of that got cheaper or less important, because the systems reading it are the same systems.

Some of it counts more than it used to. Eligibility is the big one. Google's documentation is blunt about the requirement: "To be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet, fulfilling the Search technical requirements." Read plainly, a page that is not indexed, or that has been told not to show a snippet, is invisible to the AI answer by construction. No amount of good writing fixes that, and plenty of sites are carrying that setting without knowing it.

Then there is the part nobody thinks of as SEO. For a local business, a large share of what an AI says about you never came from your website. It came from your Google Business Profile, your reviews, your listings, and the local schema markup that states your facts in a form a machine reads without guessing.

A roofing company can have the best crew in the county and still lose the answer, because the only mention of the town it actually works in sits in a footer. Our free audit will tell you which of these your site already passes.

What is genuinely new?

Four things, and they are mechanical rather than mysterious.

The result is a shortlist, and shortlists have room for three. Ranking eighth used to mean a trickle of traffic. In an answer that names three businesses, eighth means nothing arrived. This is why the work shifts toward being quotable: short, factual, self-contained statements a machine can lift without rewriting them.

One question becomes many searches. Google calls this query fan-out, which it defines as "a set of concurrent, related queries generated by the model to request more information and fetch additional relevant search results." When someone asks for a good emergency plumber in Tampa who works Sundays, the model does not look for your homepage. It runs several narrower searches and pulls the pages that answer each one cleanly. A specific page about emergency service in a specific town beats a fat homepage that mentions everything once.

There are several front doors now, and your robots.txt file decides which ones are open. OpenAI documents separate crawlers for separate jobs, and its guidance is explicit that "each setting is independent of the others." OAI-SearchBot exists to "surface websites in search results in ChatGPT's search features," and OpenAI states that "sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers." A site can be perfectly healthy on Google and locked out of an assistant entirely, with nobody on staff aware of it. The OpenAI crawler documentation lists the agents, and we walk through the file itself in our guide to robots.txt and llms.txt.

People check the answer. BrightLocal's Local Consumer Review Survey 2026, fielded with 1,002 U.S. adults, found the share of consumers using AI to find local business recommendations climbed from 6% in 2025 to 45% today, and that 97% of AI users sometimes double-check those recommendations against real reviews. This is the part that ties the new work back to the old. The AI makes the introduction, and then your reviews, your profile, and your website close it. Being named by a machine and then failing the human's five-second check is a worse outcome than not being named at all.

It is worth seeing how your own site handles these four before anyone spends money on them.

What Google says you do not need to do

Google published a mythbusting section aimed squarely at the advice going around, and it is the most useful paragraph on the internet for an owner deciding what to stop paying for. In its optimization guide, Google states: "You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search." Also: "There's no requirement to break your content into tiny pieces for AI to better understand it." Also: "You don't need to write in a specific way just for generative AI search." And on structured data: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add."

We sell schema work and crawler files, so here is the honest reading rather than a convenient one. Google is answering one narrow question: what is required to be eligible for Google's own AI features. For that question, the answer really is nothing special. Schema still does two other jobs, which Google is not ruling out. It earns rich results in ordinary search, and it states a local business's facts in a form that removes the guesswork about what the business is and is not. Crawler files are about permission rather than ranking, and permission now gets granted one assistant at a time, which is a different job from the one Google just dismissed.

The tactic that was oversold before AI and is still oversold is paying for volume of mentions. Google's own guidance puts chasing inauthentic mentions below simply publishing quality content, and the models are trained on the same web that spam has always polluted.

How do you optimize for AI search without starting over?

In order, cheapest first. This sequence is the whole recommendation.

Confirm the pages you care about are indexed and allowed to show a snippet. This is free, it takes an afternoon, and it is where most invisible businesses actually fail.
Check which assistants you have accidentally locked out in your robots.txt file.
Make the facts about your business identical everywhere a machine can read them. This is what we call entity hardening: establishing clear boundaries for what your business is and shows online, so AI tools form an accurate picture of what you do and what you do not do.
Give each real question its own page, answered in the first two sentences, with the town named.
Keep your profile current and your Google reviews coming, because that is where a local answer is assembled from.

Do all five before you buy anything with AI in its name. An agency selling you a separate AI package while your service pages are blocked from showing a snippet is selling you the roof before the walls.

One honest note on speed, because it cuts in your favor here. AI visibility moves much faster than traditional SEO. Movement is possible within 24 hours of a website change such as an updated robots.txt or llms.txt file, though results are not guaranteed. That is a different rhythm from waiting two quarters for a ranking to settle. If you would rather not work through that list yourself, our SEO, GEO and AEO work starts exactly here.

How do you know if it is working?

Three checks, in ascending cost and ascending rigor.

Ask an assistant the question your customer would ask, using your town and your trade, and read the names it gives back. Do it a few times, because answers move.

Then open Search Console. Google names its generative AI performance report as the place to see how your content is discovered through AI features on Google Search, which makes it the only first-party number available on the Google side.

Then measure properly over time. GetLocalLeads.AI, an AI visibility and digital marketing agency for local service brands, tracks citations across multiple AI platforms and delivers them in a data dashboard, which is how you tell a real change from a good day.

The honest limit: no tool sees inside every assistant, and two people asking the same question can get different names. Treat the trend as the signal and any single answer as an anecdote. Ask for the audit and you at least get the before picture in writing.

Frequently Asked Questions

Does SEO still work with AI answering the questions?

Yes. Google states its generative AI features are "rooted in our core Search ranking and quality systems," so the pages that earn AI answers are largely the pages that earn rankings. What changed is that fewer people click, so value shifts toward being named.

Is AI SEO the same thing as SEO?

Mostly. AI SEO, AEO, and GEO describe the same work aimed at answers instead of link lists. The difference is emphasis: accurate business facts everywhere, crawler permissions for more than one company's bots, and pages that answer one question.

Does schema markup help my business show up in AI answers?

Google says structured data is not required for its generative AI features and there is no special markup to add. Schema still earns rich results in ordinary search and states your business facts unambiguously, which is why we keep doing it. It is not an AI shortcut.

How long does AI search optimization take to work?

Faster than traditional SEO. Movement is possible within 24 hours of a website change such as an updated robots.txt or llms.txt file, though results are not guaranteed. Classic ranking changes still take months.

Should I hire a separate AI agency on top of my SEO company?

Usually not. One set of fundamentals serves both, and paying two firms to argue about the same website is how budgets get eaten. For a second opinion on what is broken, book a call.

The one thing worth doing this afternoon

Open an assistant, ask it the question a customer in your town would ask, and read the two or three names it hands back.

Then look at those names honestly. In most towns they are not the best contractors around. They are the ones a machine could read without guessing: consistent details, clear pages, real reviews, nothing blocked. That is a solvable problem, and it is closer to housekeeping than to magic. When you want the comprehensive version of that check, with your schema, your answer-first structure, and your crawler permissions tested for you, book a call and we will run it.

Entity SEO: Teaching AI What Your Business Is and Is Not

Brett D. Henderson on September 18, 2026

Entity SEO: Teaching AI What Your Business Is and Is Not

Entity SEO is the work of making search engines and AI assistants understand your business as one specific, identifiable company rather than a pile of keywords. It has two halves, and the second one gets skipped almost everywhere: teaching those systems what you do not do. Get both halves right and you get named for the work you want and left out of the calls you would have turned down anyway.

Key Takeaways

Entity SEO is about identity, not repetition. Search engines and AI tools store your business as a thing with attributes, not as a page that used a phrase enough times.
The boundary is the part almost nobody does. A machine needs to know where your work stops before it will confidently recommend you for the work you do.
Consistency is the whole mechanism. The same facts, in the same words, in every place a machine can read them.
Claiming more services than you deliver makes you harder to recommend for any of them.

What is entity SEO?

An entity is any distinct thing a machine can hold as one item: a person, a place, a product, a company. It has attributes (a name, an address, a phone number) and relationships (the city it serves, the trade it belongs to, the profiles that describe it). Entity SEO, sometimes written as entity based SEO, is the work of making your business one of those clearly defined items instead of a fuzzy cloud of pages.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found 14 of the 17 businesses were ranked number one by at least one engine at least once, even though the typical business appeared in only 17% of its answers.

The difference from keyword work is small to explain and large in practice. A keyword is a string somebody types. An entity is the thing they meant. When a homeowner types "emergency plumber near me," Google is not looking for pages carrying that phrase. It is looking for companies it already understands to be plumbers, open now, near that person.

In plain terms, nothing anywhere in Google or ChatGPT stores the phrase "best plumber in Sarasota." What they store is a company: a name, a category, a service area, a phone number, a set of jobs it does and a set of places other sites say it does them.

That distinction is also why entity work sits underneath answer engine optimization rather than beside it. An assistant giving one spoken answer has to be sure about the company it names, and certainty comes from a clear record, not from a well-optimized page. If you want to see what that looks like applied to a real business, GetLocalLeads.AI is happy to walk through it.

How does Google decide what your business is?

Google keeps a record of real-world things and how they connect. When it launched the Knowledge Graph in 2012, Google described it as an intelligent model "that understands real-world entities and their relationships to one another: things, not strings," and said it opened with more than 500 million objects and more than 3.5 billion facts about and relationships between them. That was fourteen years ago, and the idea has only spread further into how both search and AI assistants work.

To build the record for your company, these systems read four kinds of source: your own website, your Google Business Profile, the listings and profiles that mention you, and the structured data on your pages. Structured data is code that states your business facts in a format machines read directly. Google's Organization markup covers name, address, phone, logo and a property called sameAs, which Google defines as "the URL of a page on another website with additional information about your organization." That property is how you tell Google that the profile over there and the company over here are the same outfit.

Here is the part worth remembering. A machine does not know what you do. It knows what your sources agree you do. When they disagree, something has to break the tie, and it is usually the loudest or the oldest source rather than the correct one. The name, address and phone side of that problem, called NAP consistency, is the best-known example, and it is only one piece of it. A comprehensive AI Visibility Audit reads your website the way a machine does, if you want to see where your own record starts to disagree with itself.

Why "and is not" is the half that matters

A complete list of what you do is not a clear entity. What makes it clear is an edge, a visible place where your work stops.

Think about what an assistant is actually doing when somebody asks it who handles standing-seam metal roofs in their town. It is choosing a short list of names to say out loud. A wrong recommendation costs it more than leaving a company out, so ambiguity is a reason to skip you. A roofer whose site, profile and listings all say roofing, gutters and siding is easy to place. A roofer whose site hedges toward "general contracting, remodeling, roofing and more" has given the machine nothing firm to hold, and the machine treats that vagueness as risk.

For you, the cost lands in two directions at once. You do not get named for the work you are genuinely best at, because your record buried it under six other claims. And you do get named for work you turn away, which means you answer the phone, listen politely, and refer the caller somewhere else. That call cost you exactly what a real call costs: your attention during a working day. Run that twice a week for a year and count the working hours you gave to jobs you were never going to take.

This is what entity hardening means, and it is the reason AI visibility and brand management is built around boundaries rather than claims. The work establishes clear limits on what a business is and shows online, removing the gray areas so tools like ChatGPT and Gemini can tell the difference between where you deliver excellent service and where you simply do not offer the service. Scoping that for a specific company is what a booked call is for.

What entity confusion costs you now

This problem is not new, but it got expensive recently. BrightLocal's 2026 Local Consumer Review Survey, published in February 2026 from a poll of 1,002 US adults, found that use of ChatGPT and other generative AI tools for local recommendations rose from 6% to 45% in a single year, and that 40% of consumers trust AI platforms to give them business recommendations.

Based on one of our internal studies of 41 local service businesses, 54% wrote their business name more than one way on their own website, and 15% split one brand across more than one domain. Each variation is one more candidate record an assistant has to choose between.

That shift changes what ambiguity costs. A page of ten blue links tolerates a confusing entity, because the person reading it does the sorting. They scan, they click, they figure out that you dropped appliance repair years ago. One spoken answer does not give them that chance. Our piece on whether ChatGPT recommends businesses walks through how that retrieval works.

Take an HVAC company that stopped doing appliance repair in 2020. The service page came down, but an old listing, a directory profile and one line in the site footer never did. Six years later an assistant still reads those three sources, still associates the company with appliance work, and still sends it those calls. Nobody at the company knows why, because nothing on the website looks wrong and nothing in the reports says anything about it. The leftovers are invisible from the inside, which is why they survive. Finding yours is a good place to start a conversation with us.

How to draw the boundary where machines read it

Four places carry almost all of the weight, in this order.

Your services page. Name each service in the words your customers use, then say plainly what you do not take. "We do residential re-roofs and repairs. We do not do commercial flat roofs." Two sentences like that are worth more to a machine than a paragraph of adjectives, because they are specific, complete, and impossible to misread. Owners resist writing the second sentence. It feels like turning down money. It is the sentence that gets the first one believed.

Your Google Business Profile categories. This is where over-claiming does the most damage, and Google's own guidance is blunt about it. Google says to use as few categories as possible to describe your core business, to choose categories that are as specific as possible while still representing your main business, and not to use categories solely as keywords or to describe attributes. Google also asks businesses to reflect the business accurately and to represent it the way it is recognized in the real world, on signage and everywhere else. Two accurate categories beat nine hopeful ones. Categories are the boundary; the rest of the profile fills in the detail.

Your listings and profiles. Same company name, same phone number, same service list, same words, everywhere they appear. Not similar. Identical. Suite numbers, abbreviations and a second tracking number are the three that quietly split one company into two records.

Your structured data. Mark up what the page actually shows, and nothing else. Google's structured data policies say "Don't mark up content that is not visible to readers of the page" and "Don't use structured data to deceive or mislead users. Don't impersonate any person or organization, or misrepresent your ownership, affiliation, or primary purpose." Overstating your services in code is a policy violation with your name on it. The fields a service business needs are few, and every one of them should match what the page already says.

One rule ties all four together. Pick the words once, write them down on a single page, and then use those exact words in every one of the four places. Not a better version for the website and a shorter version for the profile. The same version. The consistency is what does the work.

What entity SEO is not

It is not schema markup with a new name. Schema is how you state your facts in a format machines read cleanly. It cannot make an inconsistent fact consistent or an untrue one true. If your profile and your site disagree, adding code to both just states the disagreement more precisely.

It is not a software purchase either. We sell this work and we will still tell you the mechanism costs attention rather than tools. Somebody has to decide what the business is, write it down, and go fix the places that say otherwise.

It is also not the first thing to fix if a business is invisible everywhere. Clarity helps a machine choose you once it can find you. If nothing about your company is indexed, if your profile is unclaimed, if there are no reviews and no pages, then coverage is the problem and clarity comes after.

And it does not mean shrinking. Adding a service you genuinely deliver is fine, and so is growing into a new trade. Claiming one you do not deliver is the part that costs you. If you are unsure which side of that line something falls on, ask.

How fast can you change what AI says about you?

Faster than traditional SEO, and unevenly. AI visibility moves much more quickly than search rankings do, and movement is possible within 24 hours of a website change such as an updated robots.txt file or an added llms.txt file (two small text files on your site that are written for crawlers rather than people), though results are not guaranteed.

The unevenness has a cause. What a model absorbed during training and what it retrieves live from the web are two different pipes. Fixing your site and your profile corrects the second one quickly. The first one updates on its own schedule, which is why the work is consistency over time rather than one afternoon of edits.

A fair way to think about the timeline: treat the first week as the live sources catching up and the following months as the older impressions fading. The practical test is not a ranking report. It is asking the assistants the questions your customers ask, in the words your customers use, and writing down what comes back. Do that before you change anything, so you have something to compare against. That is a conversation worth having before you start.

Frequently asked questions

Is entity SEO just schema markup with a new name?

No. Schema markup is one of four places your business facts live, alongside your website, your Google Business Profile and your listings. Entity SEO is the practice of making those four agree with each other. Schema is the cleanest way to state a fact to a machine, but stating a fact clearly does not help if three other sources state a different one.

What is the difference between an entity and a keyword?

A keyword is the string a person types. An entity is the thing they meant. "Emergency plumber near me" is a keyword; the plumbing company Google decides to name is an entity. Keyword work aims at matching phrases on a page. Entity work aims at being the company a search engine or an assistant already understands well enough to recommend.

Why does ChatGPT get my business information wrong?

Almost always because your sources disagree. An assistant assembles an answer from your site, your profile, your listings and whatever else mentions you. When an old directory entry lists a service you dropped or an address you left, the assistant has no way to know which version is current. It picks one. Fixing the record at the source is what changes the answer.

Do I have to list every service I offer?

List the services you want calls for, in the words customers use, and leave off the ones you take occasionally as a favor. A long list weakens every item on it, because the machine reading it cannot tell which ones are your actual business. Depth on four services beats a mention of fourteen.

Does entity SEO matter for a one-truck operation?

It matters more, not less. A single-truck company competing on clarity can beat a larger outfit whose record is a mess, because the work rewards precision rather than budget. One accurate profile, one honest services page and one consistent phone number across the web will outperform a bigger competitor claiming every trade in the county.

Start here

Pick one service you do not actually offer, or stopped offering years ago, and find every place online that still implies you do. The old listing, the stale directory profile, the line in your footer nobody has read since the site launched. That list is usually shorter than owners expect. If you would rather have someone else find them and put the record straight, that is what our SEO, GEO and AEO work is for (search engines, generative AI and answer engines), and a call is the place to start.

Voice Search Optimization: What Still Works in 2026

Brett D. Henderson on September 18, 2026

Voice Search Optimization: What Still Works in 2026

Voice search optimization in 2026 means making your business easy for an AI assistant to name out loud. It is not a separate ranking system you target. The assistants that answer spoken questions have been rebuilt on generative AI models, which means the work that gets your name said is the same work that gets you into any AI answer.

Key Takeaways

The size claims were never verified. The most quoted adoption number in this category comes from a Google survey published in 2014, and agency pages still present it as current.
The assistants were replaced. Gemini is taking over Google Assistant on speakers and displays, Alexa was rebuilt as Alexa+, and Siri now hands some questions to ChatGPT.
Google says there is nothing special to do. Its documentation states there are no additional requirements and no special markup needed to appear in its AI features.
You cannot buy your way into a featured snippet. Google's own answer to whether you can mark a page as one is "You can't."
For a local service business, your Google Business Profile answers more spoken questions than your website does.

What Is Voice Search Optimization in 2026?

Voice search optimization is the practice of making your business findable and nameable when somebody asks a device a question out loud instead of typing it. That is the whole definition. There is no separate voice index, no separate voice algorithm, and no voice ranking report that anyone can honestly send you.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found 5 of the 17 businesses were named by only one engine, or by none at all.

The vocabulary is older than the thing it now describes. When the phrase was coined, a spoken question went to an assistant that matched keywords and read one result back. Today the same spoken question goes into a generative model, which is software that composes an answer in its own words rather than reading one off a list. Same input, completely different machine on the other end. That is why this sits inside the broader category of answer engine optimization rather than beside it.

For you, that shift is good news. Nothing on this page needs a keyword tool, a developer, or a line of code you have to understand. It needs you to be clear about what your business does, where it works, and when it is open. If you would rather hand that off, GetLocalLeads.AI, an AI visibility agency for local service brands, does this work for its clients.

Was Voice Search Ever as Big as the Hype Said?

No. The behavior was real, but the numbers used to sell it were not, and a decade later nobody in the industry has gone back and corrected them.

Start with the statistic you have probably seen. The claim that 41 percent of adults use voice search every day comes from a Google-commissioned survey of 1,400 Americans aged 13 and up, published on Google's blog on October 14, 2014. That survey also found that more than half of teens used voice search daily. It was interesting in 2014. It is twelve years old now, and pages published this year still quote it without the date attached.

Then there is the prediction that half of all searches would be voice by 2020. It circulated for years, it originally covered image searches as well as spoken ones, it got attached to research firms that never published it, and it did not come true. It is still being repeated.

Here is what an independent measurement looked like instead. According to a Pew Research Center survey conducted in June 2019, one quarter of US adults said they had a smart speaker in their home, skewing younger and toward higher incomes. That figure is itself seven years old, and no newer independent equivalent turned up that was not published by a company selling marketing software. Real adoption, steadily growing, never a tidal wave.

So if you are asking whether voice search is still relevant, the fair answer is that the behavior is and the 2018 playbook is not.

What Changed When the Assistants Were Rebuilt

Every assistant that answers spoken questions has been replaced or rebuilt since 2024. This is not a refresh. The software the old tactics were aimed at is being switched off.

Google is moving Google Assistant aside. In its announcement for Gemini for Home, Google says "The Gemini for Home voice assistant will upgrade and replace the Google Assistant on your speakers and smart displays." The behavior changes with it: instead of rigid commands, you can ask a vague question, keep talking, and add exceptions partway through.

Amazon did the same thing. Announcing Alexa+ on February 26, 2025, Amazon wrote that "Conversations with Alexa+ feel expansive and natural. Whether you're speaking in half-formed thoughts, using colloquial expressions, or exploring complex ideas, Alexa+ understands what you mean."

Apple announced Apple Intelligence on June 10, 2024, giving Siri "richer language-understanding capabilities" and integrating ChatGPT access into iOS, iPadOS and macOS, with the user asked before a question is sent on.

Google also closed the door on custom voice apps. It sunset Conversational Actions on June 13, 2023, which means the third-party experiences businesses were told to build for Assistant are no longer available to users or developers at all. If you paid for one, it is gone.

Put together, the device is no longer looking your business up and reading it aloud. It is generating an answer and naming two or three companies, which is the same behavior you see when somebody asks a chatbot whether it can recommend local businesses. One short call is usually enough to find out whether the assistants currently know your business exists.

Which Old Voice Search Tactics Should You Stop Paying For?

Three of them, and in each case the platform's own documentation is the reason.

"Aim for position zero." This one was never buyable. Google's documentation on featured snippets says that "Google systems determine whether a page would make a good featured snippet for a user's search request, and if so, elevates it." Asked directly whether you can mark a page as a featured snippet, the answer Google gives is "You can't." There are ways to opt out, using the nosnippet and max-snippet rules, but there is no way to opt in. So any package priced on delivering position zero is priced on something the vendor does not control. Writing a clear answer to a real question raises your odds. Paying a retainer for a guaranteed snippet buys you a promise nobody can keep.

speakable structured data. This was the one piece of markup invented specifically for voice, and Google still labels speakable structured data "in beta and subject to change." Its scope is narrow: it "works for users in the U.S. that have Google Home devices set to English, and publishers that publish content in English," and it is framed around news content answering topical news queries. Google adds that it "does not guarantee that features that consume structured data will show up in search results." Skip this specific tag. Do not read that as a reason to skip structured data in general, because ordinary local business markup is doing real work, as covered further down.

Custom voice apps and skills. Covered above. Nobody should be commissioning one.

What Voice Search Optimization Actually Means Today

Four things hold up. If you want to optimize for voice search without wasting a quarter, this is the list, and none of it is voice-specific, which is exactly the point.

Answer real questions in plain sentences. Spoken questions are longer and more conversational than typed ones, because people talk in sentences and type in fragments. A page that states a direct answer in the first two sentences under a heading shaped like the question is a page a generative assistant can lift a sentence from. There is no trick here. It is the difference between a page that says "24/7 emergency service available" and one that says "Yes, we answer emergency calls overnight, and someone is on the road within an hour."

Be unambiguous about what you are and are not. This is what entity hardening means, which is the work of removing the gray areas so an AI tool forms an accurate picture of your business. A model that cannot tell whether you do commercial flat roofs or residential gutter repair will not put you in a shortlist of two names. Saying what you do not do is as useful as saying what you do. That clarity is the core of SEO, GEO and AEO work.

Let the crawlers in. Assistants read the open web through crawlers, and plenty of sites block them by accident. Your llms.txt and robots.txt files decide who is allowed to read your pages, and a business that has locked AI out of its own website will not be named no matter how good its copy is.

Use ordinary structured data. Local schema markup tells a machine your address, hours and services in a format it does not have to interpret. Nothing exotic is required, and Google says so directly: its documentation on AI features states that "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary," and that "There's also no special schema.org structured data that you need to add." The free AI Visibility Audit checks whether a site is set up this way.

Why Your Google Business Profile Does Most of the Work

For a local service business, the profile answers more spoken questions than the website does. When somebody asks their phone for a plumber nearby, the assistant is resolving that against local business data first. What gets sold as voice search local SEO is mostly profile accuracy wearing a more expensive name, and most voice search for local business owners comes down to the same handful of fields.

Your hours and address decide whether you are eligible at all. So do your primary category and your service area, which is the list of places you will actually drive to. An electrician whose profile says he covers one town is not a candidate when somebody two towns over asks who can come out today, no matter how good his website is. Google's own explanation of how profiles work puts it plainly: "if you own a hair salon, your business might appear in local search results for people who search for 'salons near me' or 'salons open now' because you've provided information that includes your address and hours." Stale hours mean you are invisible at the exact moment somebody with a burst pipe is asking who is open.

Consistency across the rest of the web feeds the same answer. Google builds profiles from crawled web content, licensed third-party data, contributions from users and owners, and its own interactions with the place. When those sources disagree, the machine gets less confident about you, which is why your name, address and phone number need to match everywhere they appear, and why profile optimization and ranking on Google Maps are the same project. Google Business Profile management is one of the services GetLocalLeads.AI handles, so the owner does not have to.

How Do You Know If Any of This Is Working?

Mostly you cannot measure voice separately, and that is less of a problem than it sounds.

Search Console does not split spoken queries from typed ones. The assistants do not report back at all. If someone hands you a voice search performance report, they invented the numbers, because the data to build one is not published by anyone.

So measure the thing that pays the bills. Calls and booked jobs. If your profile is accurate and your pages answer the questions people actually ask, your call volume moves whether or not anybody can prove a particular call started as a spoken question.

There is one check you can run yourself in about a minute. Ask an assistant the question one of your customers would ask, in the words they would use, and listen to which companies it names. If it names three competitors and not you, that is the same root cause as not showing up in ordinary search results, and it is worth a conversation.

Frequently Asked Questions

Is voice search optimization still worth it in 2026?

The old tactics are not worth paying for. The underlying work is, because it is the same work that gets your business named in AI answers, search results and maps listings. Treat it as one project rather than a separate line item, and judge it by calls booked. Book a call if you want someone to look at where yours stands.

What is the difference between voice search and asking an AI assistant?

Increasingly there is none. A spoken question to a modern phone or speaker is handled by a generative model that composes an answer, which is what happens when you type the same question into a chatbot. The difference is the input method, not the machine deciding which businesses to name.

Do I need speakable schema for voice search?

No. Google still lists speakable structured data as beta, limited to US users with Google Home devices set to English, and framed around news publishers. Ordinary local business structured data does far more for a service company. Add the markup that describes your address, hours and services and skip the voice-specific tag.

How do I show up for "near me" voice searches?

Claim your Google Business Profile and make sure the address, service area, categories and hours are correct, including holiday hours. Google uses that information to decide whether you are eligible for "near me" and "open now" queries. Then keep the same details consistent everywhere else the web lists you.

Can I see voice searches in Google Search Console?

No. Search Console reports queries without distinguishing how they were entered, and the assistants publish nothing. Anyone offering a voice-only breakdown is estimating. Track calls, form fills and booked jobs instead, which is what you were trying to grow anyway.

Where to Start

Open your Google Business Profile this week and check the hours and address against reality, including the holiday ones you set last December and forgot. Then read one of your service pages out loud. If it does not answer a question a customer would actually say, no assistant can quote it, and neither can a person. That is the whole of voice search SEO in 2026, and it is worth a call if you would rather not do it alone.

How to Get Cited by AI Assistants: The Checklist

Brett D. Henderson on September 18, 2026

How to Get Cited by AI Assistants: The Checklist

To get cited by AI, publish pages that answer one question directly, in sentences that stay true when they are lifted out of the paragraph, with real numbers attributed to named sources, on a page the assistants' crawlers are allowed to read. An assistant quotes a passage rather than a whole page, so the win goes to the page carrying the single cleanest liftable paragraph.

Key Takeaways

An AI citation is a sentence, not a ranking. An assistant lifts a passage from your page and names you as the source of that sentence.
Specificity is the measured lever. In the KDD 2024 study "GEO: Generative Engine Optimization", citing sources, adding quotations and adding statistics worked best, at a 30 to 40 percent relative improvement in the study's visibility measure.
Keyword stuffing does nothing here. The same study found it offers "little to no improvement."
Plain writing is now an advantage you can measure. Rewriting for fluency and readability produced a 15 to 30 percent visibility boost in the same tests.
Access comes before everything. If the crawler cannot reach your page, nothing else here matters.

What does it mean to get cited by an AI assistant?

An AI citation is when an assistant uses your page as the source for a sentence in its answer and names or links you. Someone asks ChatGPT which company to call for a burst pipe on a Sunday, and the answer comes back with three names. Being one of those three is the citation.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found Claude answered without naming a single business 11.8% of the time, 32 answers out of 272.

A ranking is a position on a list somebody scrolls. A citation is the answer itself, read by someone who never scrolls at all.

The prize moved because the customer moved. As of Pew Research Center's February 2026 survey, about half of U.S. adults now use AI chatbots, up from a third in 2024, and about four in ten use them to search for information. For local businesses the shift was faster still: BrightLocal's Local Consumer Review Survey 2026 found that use of generative AI tools for local recommendations rose from 6 percent the year before to 45 percent, putting AI third behind only Google and Facebook.

The category name for this work is answer engine optimization, which means writing a page so a machine can answer a question with it, and it is one part of AI visibility for local business. If you would rather hand that job to someone, GetLocalLeads.AI is an AI visibility agency for local service brands.

Why do AI assistants cite one page and ignore another?

An assistant does three separate things with your question, and a different quality of your page decides each one.

First it retrieves candidates. It searches, pulls a set of pages, and reads them. Being findable decides this step. Second it writes an answer, drawing sentences from whichever passages actually say something useful about the question. Being quotable decides that one. Third it attaches a source to specific claims in what it wrote. Being checkable decides that.

Step one has a hard precondition that has nothing to do with writing: the crawler has to be allowed in. Different bots do different jobs, and they are not interchangeable. OpenAI's own documentation says "OAI-SearchBot is for search" and is "used to surface websites in search results in ChatGPT's search features", while "GPTBot is used to make our generative AI foundation models more useful and safe" and crawls content that may be used in training. Blocking one is not the same decision as blocking the other. The detail of which file controls what lives in our guide to robots.txt and llms.txt.

Google sets a floor of its own. Its documentation says that to appear as a supporting link, "a page must be indexed and eligible to be shown in Google Search with a snippet", and then adds the part most agencies leave out: meeting every requirement "doesn't mean that Google will crawl, index, or serve its content." Nobody can promise you a citation. What you can do is stop disqualifying yourself.

Here is the consequence that reframes the whole job. Because the unit is a passage, a mediocre page with one excellent liftable paragraph beats a beautiful page with none. You are not competing to have the best website. You are competing to own one paragraph. The same logic explains a lot about which local businesses get recommended when a customer asks. Our team is happy to look at whether the bots can read your site at all.

What makes a source quotable?

Somebody measured this, which is rare enough in this field to be worth leaning on. The KDD 2024 paper "GEO: Generative Engine Optimization" tested nine different ways of changing a source page against a benchmark of 10,000 queries, and the three that worked best were citing sources, adding quotations and adding statistics, which produced "a relative improvement of 30-40% on the Position-Adjusted Word Count metric and 15-30% on the Subjective Impression metric." Those are the study's own measures of how much of an answer came from a given source, not traffic figures.

In our internal study of 41 local service businesses, only 20% had a dedicated FAQ page and 88% published no pricing, which left an assistant little on the page to quote.

The second finding is the one I would tape to the wall. Rewriting the source text for fluency and plain readability "also resulted in a significant visibility boost of 15-30%." Writing like a person explaining something to a customer, instead of like a brochure, is now a measured advantage rather than a matter of taste. Every owner who has ever said their website sounds nothing like them was right, and it turns out it was costing something.

In the study's live test on one engine, as of 2024, adding quotations performed best on Perplexity, at a 22 percent improvement over the baseline. One engine, one point in time, so treat it as direction rather than a promise.

Strip all of that down and three properties survive. A quotable sentence is definite: it says something rather than gesturing at something. It is specific: there is a number, a date, a name or a boundary inside it. And it is self-contained: lift it out of the paragraph, print it on its own, and it is still true and still makes sense.

Most service-business web copy fails all three at once, and it fails them politely. "We pride ourselves on quality workmanship" is not wrong. It just has nothing in it for a machine to quote. Our comprehensive AI Visibility Audit checks a site against this kind of standard if you would rather see it scored than guess.

The citable-content checklist

Pick one page. Not your homepage, which tries to do everything, but the page that answers the question your best customers actually ask. Run these ten items against it and mark each one pass or fail.

Does the page answer one question, not five? A page covering five topics has no clean passage about any of them, and the assistant goes elsewhere for each one.
Is the answer in the first 50 words, above any setup? The first coherent answer block is what gets extracted. A page that warms up for three paragraphs loses to one that does not.
Does the heading match how a customer would say it out loud? People prompt in questions. "How long does a furnace last" beats "Equipment Lifecycle Considerations" every time.
Can you lift any single sentence out and have it still be true? Test it literally. Copy a sentence into a blank document. If it needs the two sentences before it, it cannot be quoted.
Is there a real number on the page, and does it say where the number came from? A figure with a named source behind it is the kind of sentence the KDD 2024 study found assistants pick up most readily.
Is there a date on the page, and on any claim that could age? "As of 2026" tells a machine the page is current and tells a reader you are not hiding a five-year-old figure.
Does a named human appear as the author, with somewhere to go and check who they are? Google's own guidance asks "Is it self-evident to your visitors who authored your content?" and whether "pages carry a byline, where one might be expected".
Do you call the same thing the same name every time? Including your own business name, your service names and your service area. Two spellings read as two different companies.
Is the page reachable by the assistants' crawlers and indexed by Google? Both, separately. Passing one and failing the other still ends in silence.
Would a competitor's page honestly be a better source for this question than yours? If the answer is yes, fix that before anything else on this list.

Score it out of ten. Anything below seven and the page is not a citation candidate yet, whatever the traffic says.

How do you rewrite a sentence so it survives being lifted?

Put something in the sentence that is only true of you: a number, a place, a method or a boundary. The fastest way to see the difference is side by side, and these are illustrative rewrites, not claims about any particular company.

A plumbing company writes: "We pride ourselves on fast response times." The rewrite: "We answer emergency calls within two hours in the counties we serve, measured on our call tracking system." What changed is that the second version contains a commitment and a method, so it can be quoted and checked.

A roofer writes: "Serving the greater metro area for years." The rewrite: "Founded in 2009, we work in Jefferson, Shelby and St. Clair counties." Two vague words became a year and three place names, which is exactly the kind of detail an assistant needs to decide whether to name you for a local question.

An HVAC company writes: "We handle all types of heating systems." The rewrite: "We service gas furnaces, heat pumps and mini-split systems, and we do not work on oil-fired boilers." Naming what you do not do feels like giving something up. It is the opposite. A boundary is information, and information is quotable.

The rule underneath all three: if a sentence would be equally true printed on your competitor's website, it will never be quoted, because it carries nothing. Saying plainly what a business does and does not do is a discipline of its own, and it is worth a longer conversation than this piece can give it. We rewrite pages to this standard as a matter of course.

What does not work, and what Google actually says

Keyword stuffing does not work. The GEO study tested it directly and found it offers "little to no improvement" on what generative engines produce. Repeating "emergency plumber Birmingham" eleven times moves nothing.

Neither does a magic file. Google's own documentation is unusually blunt about this: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary", and "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add."

That deserves an honest nuance rather than a victory lap. Structured data is still worth having, because local schema markup removes ambiguity about who you are, where you work and what you sell for every machine reading the page. That is a real benefit. It is simply a different argument from "Google requires it," and anyone selling you the second one is selling you something.

The limit that matters most is the plainest. Being quotable does not help if the page has nothing worth quoting. Formatting is packaging. If the underlying page has no numbers, no boundaries and no point of view, structuring it better just makes the emptiness easier to scan.

Getting cited by ChatGPT is also not quite the same job as getting cited by Perplexity, and Google's AI Overviews work differently again. Those are their own subjects. If you want the short version for your own site, we can walk you through it.

How do you know whether any of it worked?

Not by watching traffic. The scoreboard that works is a fixed set of customer questions, run through the assistants each month and scored on whether you are named.

Traffic misleads because the answer is increasingly the destination. Pew Research Center's analysis of browsing data from March 2025, published in July 2025, found that Google users who saw an AI summary clicked a traditional search result in 8 percent of visits, against 15 percent for users who saw no summary, and clicked a link inside the summary itself in just 1 percent of visits. Waiting for a traffic line to move is waiting for the wrong number.

Building that scoreboard takes an afternoon. Write down the ten or fifteen questions a real customer would type before hiring someone in your trade and your town. Run the same list, word for word, across the assistants on the same day each month, and record whether you are named. That is a score you can actually hold someone to.

GetLocalLeads.AI measures AI citations across multiple platforms and reports them in a live dashboard, which is the same measurement done continuously rather than by hand. Movement can be quick: a website change can show up in AI answers within 24 hours, though results are never guaranteed. You can see what that looks like in practice in our AI visibility case study, or ask us about AI visibility and brand management directly.

Frequently asked questions

Does ranking number one on Google get you cited by AI?

Not by itself. Ranking and citation are decided by different systems for different jobs. Strong rankings help you get retrieved in the first place, but what gets quoted is the passage that answers the question most cleanly. A page at position six with a sharp, self-contained answer regularly beats the page at position one that buries it.

Do I need schema markup to get cited by AI?

Google's documentation says no: "There's also no special schema.org structured data that you need to add" to appear in its AI features. Schema is still worth having: it removes ambiguity about your name, location and services for every system reading your site. Treat it as clarity work, not a requirement someone can charge a premium for.

How long does it take to get cited by AI assistants?

Faster than traditional SEO, which is the good news here. Movement is possible within 24 hours of website changes such as an updated robots.txt file, though results are not guaranteed and a single fast win is not a trend. Judge it over two or three monthly prompt runs rather than a week.

Is this just SEO with a new name?

Mostly it overlaps, and anyone calling it a brand new discipline is selling a rebrand. The part that is genuinely different is the target. Classic SEO optimizes for a click on a list. AI citation optimization writes for extraction, where the sentence itself is the product and nobody may ever visit the page it came from.

How do I check whether AI is citing my business right now?

Open ChatGPT, Gemini and Perplexity, and ask each the way a customer would: "who is the best roofer in [your town]". Ask five or six variations and write down every business named. Do it again next month with the identical prompts. If you would rather have that run and scored for you, that is what our free AI Visibility Audit is for.

Where to start this week

Pick the one page that answers the question your best customers actually ask. Run the ten items against it, then fix the three that failed hardest. That is a morning's work, and it will teach you more about how your site reads to a machine than any report will.

The businesses getting named in AI answers right now are rarely the biggest ones in town. They are the ones whose pages say something specific enough to repeat. When you want help making yours say it, book a call.

How to Appear in AI Overviews as a Local Business

Brett D. Henderson on September 18, 2026

How to Appear in AI Overviews as a Local Business

To appear in AI Overviews, your page has to be indexed by Google and allowed to show a snippet in normal search results, and it has to be one of the pages Google's ranking systems already trust for that question. There is no separate submission, no special file, and no markup that buys you a spot. Google says so in its own documentation. Everything else you have been sold about AI Overview optimization sits on top of those two conditions.

Key Takeaways

Google publishes one hard rule: a page must be indexed and eligible to be shown in Google Search with a snippet before it can appear in any generative AI feature on Search.
AI Overviews are grounded in Google's core Search ranking systems, so presence in the box follows the organic trust you have already earned.
Google's own guide lists llms.txt files, special schema markup and rewriting for AI under what you do not need to do. Most "AI Overview ranking factor" lists sell you those anyway.
For local searches the AI Overview is one box among several. On straight "near me" queries the local pack still does most of the work.

What is a Google AI Overview, and why should an owner care?

A Google AI Overview is the AI-written summary that sits at the top of the results page, above the blue links, with source links attached to it. A customer types a question, Google writes a few sentences back, and names the handful of sites it drew from. Sometimes a business gets named in that summary. Usually it does not.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found only 7 of the 17 businesses were named at least once by all five engines.

The reason to care is volume. Google said at its I/O event on 19 May 2026 that AI Overviews now has over 2.5 billion monthly active users. In a separate announcement on 3 June 2026, it said AI Mode had passed one billion monthly users. Those are Google's own numbers, not an agency's estimate.

Two quick boundaries so the rest of this makes sense. AI Mode is the separate conversational tab inside Google, and it behaves differently. ChatGPT, Perplexity, Gemini and Claude are different products entirely, with their own rules about who gets named. This piece is about the Google box that appears on the regular results page and nothing else.

If you want to know where your business stands across all of it, that is what a conversation with GetLocalLeads.AI is for.

How does Google decide which sites go into an AI Overview?

It starts with a gate you can check today. Google's guide to optimizing for generative AI features, last updated on 10 July 2026, states it directly: "To be eligible to be shown in generative AI features on Google Search, a page must be indexed and eligible to be shown in Google Search with a snippet, fulfilling the Search technical requirements." Fail that and nothing else you do matters.

Past the gate, the mechanism is less mysterious than the marketing around it. Google describes the technique behind these answers as grounding, which it defines as a way "to improve the quality, accuracy, and freshness of AI responses by relying on our core Search ranking systems." Read that again slowly. The AI is not running a second, secret contest with its own winners. It is summarising pages that the existing ranking system already surfaced, then linking back to them.

There is one wrinkle worth knowing. Google says both AI Overviews and AI Mode may use a technique it calls query fan-out, which it defines as "a set of concurrent, related queries generated by the model to request more information and fetch additional relevant search results." In plain terms, one typed question can trigger several searches behind the scenes. That is why a page sometimes gets cited for a question it never targeted by name, and why covering a subject properly beats matching a phrase exactly.

So here is the sentence to take away: how to rank in AI Overviews is really a question about how well you rank already. Presence in the box sits downstream of the trust you have built in ordinary search. That is good news for an owner with a real business and bad news for anyone selling a shortcut.

Our free AI Visibility Audit reads a site the way those systems do.

Why isn't my business in the AI Overview?

Almost always one of five things, and two of them are switches somebody flipped rather than anything about your content.

First, the page may not be indexed at all, or it may carry a noindex tag from a redesign nobody cleaned up. Until Google has the page in its index, nothing downstream can happen. This is the same root cause behind most cases of a business not showing up on Google at all.

Second, snippet controls. Google names nosnippet, data-nosnippet, max-snippet and noindex as the controls that limit what it can show from your pages in Search. A developer who added them to stop scrapers also stopped Google from quoting you.

Third, and this one is new enough that most owners have never heard of it: on 3 June 2026 Google announced a Search Console setting where "website owners can decide if they want their site to appear in and help ground responses in our generative AI Search features." Sites using that control "will not receive traffic or impressions from our generative AI features." If anyone touched your Search Console settings this year, check it.

Fourth, your page simply is not one Google trusts for that question yet. That is a ranking problem wearing an AI costume.

Fifth, for local questions specifically, your Google Business Profile may be wrong, thin, or eighteen months stale.

If you would rather somebody else ran those five checks, book a call and we will run them with you.

What Google says you do not need to do

Five things are being sold as AI Overview requirements that Google's own guide lists under what you do not need to do.

Special files, including llms.txt. Google's wording is blunt: "You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search." That file is a live question for other assistants, but as a Google AI Overview lever it is not one.

Structured data. Google states that "structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." Do not strip your schema out over that. Structured data is the standard code that tells a machine your phone number is a phone number, and it still earns rich results and still helps software read your site. It is simply not the secret key it gets sold as.

Chopping your content into bite-sized chunks. Google says there is "no requirement to break your content into tiny pieces for AI to better understand it."

Rewriting your site in some special AI-friendly voice. Google says you "don't need to write in a specific way just for generative AI search."

Buying mentions on other sites to make the models notice you. Google names seeking inauthentic mentions as unhelpful.

Its AI features documentation puts the whole thing in one line: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." When a vendor's pitch is built on a lever the platform has publicly denied, you have learned something about what you are buying.

How much does the AI Overview actually matter for a plumber or a roofer?

It matters, and it matters less than the headlines suggest, and the difference depends entirely on what your customer typed.

A 2025 study of 540 local searches, collected by hand across three US cities and six trades, found AI Overviews on only 15% of straight local-intent searches, while the local pack showed up on 93% of them. Flip to informational searches and the numbers invert: AI Overviews on 92%, local packs on 6%. That data is from May 2025 and the mix has almost certainly moved since, but the shape is the useful part. Someone typing "plumber near me" mostly sees a map and three listings. Someone typing "how much does it cost to replace a water heater in Denver" mostly sees an AI summary.

Then there is what people do with the box once it appears. Pew Research Center tracked the browsing of 900 US adults across 68,879 Google searches in March 2025 and published the results that July. When an AI summary appeared, people clicked a traditional search result on 8% of visits. When no summary appeared, they clicked on 15%. Clicking a link inside the summary itself happened on 1% of visits.

Here is how to hold both facts at once. The AI Overview is increasingly where the research happens, and the local pack is still where the call happens. The same underlying work feeds both. GetLocalLeads.AI is an AI visibility and digital marketing agency for local and multi-location service brands, and it exists to make that work add up to phone calls rather than screenshots.

What can a local business owner actually control?

Four things, and you already own all of them.

Based on one of our internal studies of 41 local service businesses, 54% had no FAQ content anywhere, so the questions buyers ask most had no answer on the page for Google to draw from.

Indexation and snippet eligibility. Open Google Search Console, run the URL Inspection tool on your main service pages, and confirm each one is indexed. Then check that no stray noindex or nosnippet directive is sitting in your page code. This is the only hard gate Google publishes, and fixing it costs nothing but an afternoon.

A real answer in the first two sentences. Under each heading on your site, answer the question directly before you build up to it. Not because AI needs a special format (Google has said it does not), but because a self-contained answer is the thing any summariser can lift cleanly. Write it about your market, with your price ranges, your permit rules, your response times. Generic text gets skipped.

Your Google Business Profile. Google's guide says outright that "using products like Merchant Center and Google Business Profiles can help your products and services to be visible in both AI responses and other Google Search results." That is Google naming the asset you already have. Hours, services, service area and categories need to be current, and our guide to Google Business Profile optimization covers the fields that matter most.

One consistent story about what you are. Entity hardening means making it unambiguous what your business does, where it works, and what it does not offer, so the machines stop guessing. If your site says commercial HVAC, your profile says residential, and an old directory lists you as a general contractor, you have given three answers to one question. That is the core of AI visibility management, and it runs alongside the SEO, GEO and AEO work rather than replacing it.

None of this is fast. That is the honest trade-off, and also the argument for doing it: the same work serves ordinary search, where most of your calls still come from. Get My Audit if you want a read on how much of it is already in place.

How do you check whether you appear in AI Overviews?

Two ways, one free and one manual.

Google now reports this directly. Its guide says to "use the Generative AI performance report in Search Console" to measure how your content is performing in generative AI features on Search and Discover. It will not tell you which sentence got quoted, but it will tell you whether the traffic exists.

The manual check takes five minutes. Write down the five questions a customer would actually type before hiring you, run them in an incognito window, and note who gets named in the summary. Do it again next month.

One caution: these answers are personalised and they shift by location, device and phrasing. A single check is an anecdote. A fixed set of questions, run on a schedule, is data, which is the difference between guessing and knowing. The manual check is covered in more depth in our guide to checking whether AI recommends your business.

We track this across a fixed prompt set for the brands we work with, and a call is the quickest way to see what that looks like.

Frequently asked questions

Do I need an llms.txt file to appear in Google AI Overviews?

No. Google's documentation says you do not need new machine-readable files, AI text files, markup or Markdown to appear in Google Search. The file is a separate conversation for other assistants, but Google does not read it for this.

Does blocking AI crawlers remove my business from Google AI Overviews?

Google-Extended is the control publishers use to manage whether Google may use crawled content for training and grounding in some of its other systems. Google's crawler documentation states that Google-Extended "does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search." It is also not a separate crawler with its own user agent.

Do I need schema markup to rank in AI Overviews?

Schema markup is structured code that labels the facts on your page so software can read them. Google says structured data is not required for generative AI search and there is no special schema.org markup you need to add. Keep yours anyway, because it still earns rich results.

Can I see AI Overview traffic in Google Search Console?

Yes. Google directs site owners to the Generative AI performance report in Search Console to measure how content is performing in generative AI features across Search and Discover. It shows you whether clicks and impressions are arriving. It does not show which passage was used.

My competitor is in the AI Overview and I am not. What is different about their site?

Usually three things. They already rank for that question in ordinary search, their page answers it plainly in the first two sentences, and their business details match everywhere Google checks. If you want to know which of those is costing you, a call is the fastest way to find out.

The work is the same work

Getting a local business into a Google AI Overview takes the same checklist as AI visibility for local business everywhere else, the one that gets you named when somebody asks who to call. Be findable, be quotable, be consistent about what you are. That is why the work is worth doing, and it is also why buying a product aimed at this one box is a poor use of an owner's money. The box will change. It looked different a year ago and it will look different a year from now. Being the obvious answer to a customer's question will not go out of date. Book a call with GetLocalLeads.AI when you want that handled properly.

How to Optimize for Perplexity, Gemini and Claude

Brett D. Henderson on September 18, 2026

How to Optimize for Perplexity, Gemini and Claude

To optimize for Perplexity, Gemini and Claude, do the same work that gets you named in
ChatGPT, then check one thing per assistant: whether the specific crawler it sends is
allowed to read your site. Each company runs two or three separately named bots with
separate jobs, and your robots.txt file treats every one of them as a different visitor. A
rule written to keep your content out of AI training can quietly keep your business out of
AI answers, and most owners never look.

Key Takeaways

Optimizing for Perplexity, Gemini and Claude is mostly the same job as ChatGPT. They
differ in how they reach your site and whether the user sees your name, not in what they
reward.
Each company runs several crawlers with different jobs. OpenAI documents four,
Anthropic three, Perplexity two. Blocking the wrong one removes you from answers.
Google-Extended does not control Google Search. Google states it governs training and
grounding for Gemini models, and is not used as a ranking signal in Google Search.
None of the four publish how they score sources. Any article quoting you a citation
percentage or a freshness weight is quoting a guess.

Who is actually using the other assistants?

In a Pew Research Center survey fielded in February 2026, about half of U.S. adults said
they use AI chatbots. ChatGPT led at 44% of adults. Gemini reached 24%, Copilot 17%, Meta
AI 14%, Grok 8% and Claude 6%. Those figures come from
Pew's 2026 report
on American AI use.

In our AI gap report analysis of 17 local businesses across 1,352 AI answers, we found when at least one engine named a business, every engine agreed only 9% of the time, 10 combinations out of 110.

Read that as an owner rather than a marketer. ChatGPT is the biggest room, and it is not
the only room. Roughly one in four adults is asking Gemini. When a homeowner with a failing
water heater asks any of them who to call, the answer comes back as two or three names, not
a page of ten blue links. Being absent from one assistant is not a rounding error when the
shortlist is that short. Our guide to
AI visibility for local business
covers how that shortlist gets built.

The assumption worth retiring is the one most owners make after a good result: they type
their trade and their city into ChatGPT, see their own name, and treat the job as finished.
That test covers one assistant on one day. It says nothing about whether Perplexity can
reach the site at all.

If you want a second opinion on where your business stands, that is what a conversation
with us usually starts with.

How does each assistant actually find your website?

A crawler is an automated visitor that reads your pages. These four companies send crawlers
to do three different jobs, and the difference matters because you can allow one and refuse
another.

The first is training, collecting text that may be used to build future versions of the
model. The second is search indexing, building the catalog the assistant searches when
someone asks a question. The third is a live fetch, when a person asks something right
now and the assistant reads a page to answer. Blocking one job does not block the others.

Here is what each company documents about itself, as of September 2026.

Assistant	Crawlers it sends	How it reaches your page	What controls access	How your name shows up
ChatGPT	GPTBot, OAI-SearchBot, ChatGPT-User, OAI-AdsBot	OAI-SearchBot surfaces sites in ChatGPT's search features; ChatGPT-User visits when a person asks	robots.txt, per bot. Opting out of OAI-SearchBot removes a site from ChatGPT search answers	Linked sources in the answer
Perplexity	PerplexityBot, Perplexity-User	PerplexityBot surfaces and links sites in Perplexity's search results; Perplexity-User visits on a user's request	robots.txt for PerplexityBot; Perplexity-User generally ignores it because a person triggered the visit	A link to the page included in the response
Gemini and Google AI surfaces	Googlebot, Google-Extended	Google's search index. A page must be indexed and eligible to show with a snippet	robots.txt for Googlebot, plus nosnippet, data-nosnippet, max-snippet and noindex. Google-Extended covers training and grounding only	Supporting links beside the response
Claude	ClaudeBot, Claude-SearchBot, Claude-User	Claude decides per question whether to search. Claude-SearchBot works on search quality; Claude-User visits when a person asks	robots.txt, honored per bot	Citations carrying the source URL, page title and a short quoted passage

Run your eye down the fourth column. That is the only one you control directly this week,
and it is where nearly every self-inflicted visibility problem lives. A developer who added
a broad AI block in 2024, or a security plugin that refuses unfamiliar user agents, can take
a business out of three assistants without anyone noticing. Nothing breaks. No error
appears. The phone is quieter than it was, and the cause is a text file almost nobody opens.

One caution before you read further on this topic elsewhere. None of the four publish how
they score or order the sources they pick, so the table is the documented part.

What does optimizing for Perplexity actually require?

Optimizing for Perplexity starts with the most useful sentence Perplexity has published
about itself. PerplexityBot, its documentation says, is "designed to surface and link
websites in search results on Perplexity. It is not used to crawl content for AI foundation
models." Read that twice if you have ever been advised to block AI bots. A rule aimed at
keeping your content out of AI training does not remove you from Perplexity. A rule aimed at
PerplexityBot does, and it is the rule most blanket blocks accidentally write.
Perplexity's crawler documentation names both
agents.

The second agent is Perplexity-User, which supports actions inside Perplexity. When somebody
asks a question, it may visit a page to help answer it. Because a human triggered that
visit, the documentation says the fetcher generally ignores robots.txt rules, and Perplexity
says it will include a link to the page in its response. A user-initiated visit can surface
your name even where an indexing rule would not.

Perplexity also publishes the IP address ranges both agents use. That matters more than it
sounds. Anything can claim to be PerplexityBot in a server log, and the published ranges are
how you tell a real visit from a bot wearing someone else's name. If your host has ever
blocked "suspicious" traffic in bulk, this is where you check what got caught.

Access questions like this are the fastest part of our free audit to answer.

What does Gemini do differently?

Gemini is the cheapest of the four to cover, because you are probably already doing the
work. Google's guidance on its
AI features states that
"to be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be
indexed and eligible to be shown in Google Search with a snippet," and that there are "no
additional requirements … nor other special optimizations necessary." Ordinary indexing
plus snippet eligibility is the entry ticket, though Google adds that meeting every
requirement still does not guarantee it will serve your content.

The confusion worth clearing up is Google-Extended. It reads like a switch that hides you
from Google's AI, and it is not. Google's
crawler documentation
describes it as a control for whether crawled content may be used for training future Gemini
models and for grounding, then states that "Google-Extended does not impact a site's
inclusion in Google Search nor is it used as a ranking signal in Google Search." Turning it
on is a decision about model training, and treating it as a way to manage how you appear in
Google's answer surfaces leads owners to fix the wrong file.

So what people call Gemini SEO is, for a local business, mostly the Google work already in
progress: get indexed, stay snippet-eligible, keep the facts about your business correct.

Worth a short call if you are unsure which of those three you are missing.

What does Claude do differently?

Claude has the most distinctive behavior of the four, and it has nothing to do with content
format. Anthropic
documents three crawlers:
ClaudeBot for training, Claude-SearchBot for
search result quality, and Claude-User for visiting sites when a person asks a question. Its
bots honor standard robots.txt directives. The interesting part comes earlier. In Anthropic's
web search documentation,
Claude "determines when to search based on the prompt," searching when the request depends
on information that is current or changing, and answering directly when the request draws
on stable knowledge.

For a local business, that split usually falls your way. Which roofer to call in a specific
town in 2026 is current information, so the search happens. And when it does, citations are
always enabled, each one carrying the source URL, the page title and a short quoted passage.
Your name arrives attached to a link rather than absorbed anonymously into a paragraph.

The part that should give you pause is the other half. When an assistant answers from what
it already believes rather than from a fresh search, the accuracy of that belief is the
whole game. Entity hardening, which means removing the gray areas so a machine can tell
exactly what your business is and what it does not do, is the work that fixes it.

Ask us about that clarity; most owners have never had anyone build it.

What works the same across all four?

Here is the finding the agency internet keeps burying. The four assistants converge much
more than they diverge. All of them want a page they can reach, an answer near the top of
it, and a business they can identify with confidence. Nobody publishes a preference for a
listicle over an essay, or for 1,200 words over 2,000. Those claims come from vendors
selling a tracking subscription, not from the companies building the systems.

Four things pay on every one of them. Let the right bots in, which means reading your own
robots.txt instead of assuming. Answer the question in the first paragraph of the page,
because a machine lifting an answer takes the first complete one it finds. Write facts as
standalone sentences that survive being quoted without the paragraph around them, since
that is literally what a citation is. And describe your business identically everywhere it
appears online, so the same name, address, services and service area show up in every
place a model might check. That last one is the least exciting and does the most work.

People call this Perplexity SEO, Gemini SEO or AI assistant SEO depending on who is
selling. The vocabulary is unsettled and the underlying job is not. Anyone quoting you four
separate content strategies is quoting three you do not need. One well-built site, checked
per assistant for access, covers the field.

That combination of access work and entity clarity is what the GetLocalLeads.AI
AI visibility and brand management
service exists to handle.

How do you check whether each assistant can see you?

Open yourdomain.com/robots.txt in a browser. Read it yourself, even if you do not consider
yourself technical, because you are looking for names rather than syntax. The ones that
matter are GPTBot, OAI-SearchBot and ChatGPT-User for OpenAI, PerplexityBot and
Perplexity-User for Perplexity, ClaudeBot, Claude-User and Claude-SearchBot for Anthropic,
and Google-Extended for Gemini training. A Disallow line under any of those names is a
decision somebody made on your behalf. Our explainer on
robots.txt and llms.txt covers what
belongs in the file.

Then run the test that costs nothing. Ask each of the four assistants the question your
customer would ask, in your own city, and write down the names that come back. Four rows,
one per assistant, plus a column for whether you appeared at all and a column for who
appeared instead. The gaps in that little table tell you where to look first, and they are
usually not where owners expect. An owner who is named in ChatGPT and missing from
Perplexity has an access problem, not a content problem.

Changes here can move quickly. AI visibility moves faster than traditional SEO, and
movement is possible within 24 hours of a website change such as an updated robots.txt or
llms.txt file, though results are not guaranteed. Our
eight-week case study
shows what a fuller push looks like.

The done-for-you version of that check is our free AI visibility audit.

Frequently asked questions

Does Perplexity use Google's index?

Perplexity runs its own crawler, PerplexityBot, whose documented job is surfacing and
linking websites in Perplexity's search results, and which is explicitly not used to train
foundation models. The documentation stops there, so anyone naming the index underneath is
guessing. People ask how to rank on Perplexity, which is the wrong verb: it cites sources
rather than ranking them.

Should I block AI crawlers from my website?

Rarely, and never in bulk. All four separate training from search, so you can refuse one and
allow the other. OpenAI's documentation notes that a webmaster can allow
OAI-SearchBot while disallowing GPTBot, and that sites opted out of OAI-SearchBot "will not
be shown in ChatGPT search answers." A blanket block removes a business from answers while
failing at its purpose.

Is Gemini the same as Google AI Overviews?

Related but not identical. Gemini models power Google's AI features, and Google documents
one eligibility rule for those surfaces: a page must be indexed and eligible to show with a
snippet in Google Search. Google-Extended is a training control and does not affect Search
inclusion.

Does Claude actually read my website?

It can. Anthropic runs Claude-User, which visits sites when a person asks, and
Claude-SearchBot, which works on search quality. Claude decides per question whether to
search, answering from existing knowledge on stable topics and searching when the
information is current, as it is for which local business to hire.

Which AI assistant should a local business focus on first?

Start where the volume is: by Pew's February 2026 numbers that is ChatGPT at 44% of U.S.
adults and Gemini at 24%. Sequence matters less than people think, because the work that
earns a mention in one earns it in the others. Get your access right and your business
described consistently, then verify each one. If you would rather someone ran it, book a
call.

Where to start this week

Do the two cheap things first. Read your robots.txt and note every AI user agent named in
it, then ask all four assistants your customer's question and record who they name. Most
owners find a surprise in the first task and an uncomfortable answer in the second.
Perplexity optimization, and the same work for Gemini and Claude, rarely starts with writing
more content. It starts with finding out what you have already been blocking. When you want
the whole picture rather than two data points, we will handle the brand management
side for you.

AI Visibility Statistics From 1,352 Real AI Answers

How we measured this

How often does AI recommend local businesses?

The same business, different city, different answer

Not every engine answers the same way

Being named is a switch, not a ladder

Why position four does not exist

How this compares to the bigger studies

What these numbers cannot tell you

Frequently asked questions

How often does AI recommend local businesses?

Why doesn't ChatGPT recommend my business?

Which AI engine recommends local businesses most often?

Is being named second or third good enough?

How do I check whether AI recommends my business?

Run it on your own business this week

AI Visibility Audit: What 41 Local Business Sites Revealed

Key Takeaways

What does an AI visibility audit check?

Every business explained its services. None scored above 41.

How many businesses wrote their own name more than one way?

How common were typos and leftover template text?

What happens when a website contradicts itself?

Why does so much proof look borrowed?

Which answers does AI look for that these sites left out?

What should a local business fix first?

How we ran these audits, and what the scan cannot see

Frequently asked questions

What does an AI visibility audit check?

Do I need an AI visibility audit if I already rank on Google?

What is a good AI visibility score?

How much does an AI visibility audit cost?

Can I check my own website for these problems?

More GetLocalLeads.AI case studies

Is SEO Dead? What AI Answers Really Changed

Key Takeaways

What is the strongest case that SEO is dead?

What the click data is actually measuring

Will AI replace SEO?

Is local SEO dead?

What actually died

When SEO is genuinely the wrong spend

Does SEO still work for a small business?

Frequently Asked Questions

Is SEO dead in 2026?

Will AI replace SEO?

Is local SEO dead?

Is SEO worth it for a small business?

How long does search work take to show results?

What to check before you cancel anything

ChatGPT SEO: How to Rank on ChatGPT

Key Takeaways

Why Your Customers Are Asking ChatGPT Before They Call You

What ChatGPT SEO Actually Means

Step One: Make Sure ChatGPT Is Allowed to Read Your Site

Step Two: Write Pages a Machine Can Quote

Step Three: Make Every Mention of You Say the Same Thing

Step Four: Test It With the Prompts Your Customers Actually Use

How Long This Takes, and What It Will Not Do

Frequently Asked Questions

Does ChatGPT use Bing's index?

Can I pay to have ChatGPT recommend my business?

How long does ChatGPT SEO take to work?

Will blocking AI training hurt my ChatGPT visibility?

Is ChatGPT SEO the same as ranking on Google?

One Last Thing

AI Search Optimization: What Changes and What Does Not

Key Takeaways

What is AI search optimization?

What actually changed?

Which parts of your SEO still count?

What is genuinely new?

What Google says you do not need to do

How do you optimize for AI search without starting over?

How do you know if it is working?

Frequently Asked Questions

Does SEO still work with AI answering the questions?

Is AI SEO the same thing as SEO?

Does schema markup help my business show up in AI answers?

How long does AI search optimization take to work?