AI's Impact on Local SEO and Local Queries

99Visibility Research · Updated October 21, 2026 · 9 min read

Short answer

AI answers change local SEO by putting a model between the searcher and the list of businesses. The model rewrites the question, fetches some pages, and names a few businesses. You can't control which ones it picks. You can control two things: whether AI crawlers can reach and index your site, and whether your business facts are stated clearly, consistently and correctly on pages they can fetch.

Key takeaways

A local AI answer is built from a model's memory, the pages it retrieves for that query, or both, and you only influence the retrieved part.
Crawler access is the first gate: robots.txt rules are per bot, and a firewall or CDN can block a crawler that robots.txt allows.
Most non-Google AI crawlers don't run JavaScript, so hours, prices and addresses loaded by scripts may be invisible to them.
State your address, hours, services and prices as plain text on fetchable pages, and keep them identical everywhere you publish them.
No markup, phrasing or page template has credible evidence behind it as a way to get named in AI answers.

How an AI answer handles a local query

Say you run a dental clinic in Austin. A prospective patient asks an AI assistant: "Which dentists in Austin take emergency appointments on Saturdays?" This is a hypothetical, but the steps are documented by the engines themselves.

First, the engine decides whether to search the web at all. As of October 2026, retrieval is per query and discretionary on every major platform. ChatGPT searches at the model's discretion. Claude's web search is a user toggle. Google AI Overviews don't appear on every search, and they vary by location and language (see whether AI Overviews differ by location and language).

If the engine doesn't search, it answers from what the model learned in training. That knowledge is frozen at a cutoff. OpenAI lists a February 16, 2026 cutoff for its GPT-5.6 models and August 31, 2025 for the Instant model behind chat-latest. If your clinic moved in March 2026, a no-search answer may still give the old address.

If the engine does search, it rarely uses the patient's words as typed. OpenAI says ChatGPT search "rewrites your query into one or more targeted queries" and sends them to search providers. Google describes a "query fan-out" that issues several related searches across subtopics. So one local question can become several searches: Saturday hours, emergency dental care, Austin dentists, reviews.

The engine then reads a set of retrieved pages and names a few businesses. Those pages may be your site, or directories, review sites and news articles that mention you. What matters next is whether your clinic appears in that set, and whether the facts the engine reads about you are right.

The four stages, applied to a local business

It helps to separate four stages, because each one has a different owner.

Parametric (memory). What the model knows with no web access. Nobody changes it this quarter. Facts published after the cutoff simply aren't there.
Access. Whether a crawler may fetch and store your pages. You and your hosting or CDN provider control this.
Retrieval. Whether your page enters the model's context for a given query. The engine controls this, influenced by how relevant your page is to the searches it runs.
Generation. Which retrieved pages the model uses and cites. The model controls this.

Almost every published "GEO technique" was measured at the generation stage, on pages already inside the model's context. A July 2026 critical survey of 45 GEO studies concluded that no reviewed technique shows a stable, cross-platform causal effect on discoverability. For a local business, that means rewriting your service pages in a special style is not a proven route into the answer. Being reachable, indexed and factually clear is the part with documented mechanisms. The wider picture is in how AI search features affect organic results.

What is and isn't in your control

Stage	What decides it	Your control
Parametric memory	Training data and the model's cutoff date	None in the short term
Access	robots.txt, firewall or CDN rules, server-side rendering, index and snippet eligibility	High. These are checkable and fixable
Retrieval	The engine's rewritten queries and its index	Partial. Topical relevance and being in the index the engine uses
Generation	The model's choice among retrieved pages	None directly. Accurate, quotable facts help the engine say the right thing

Two honest limits apply to local businesses in particular. First, when an engine cites directories and review sites instead of business websites, being named accurately on those sites is associated with appearing in the answer. That's an association in the evidence, not a promised effect. Second, nothing published measures how often a specific business gets named for a specific local question. Any single score averages answers given with and without live search.

Index access: the first gate

Access problems remove you silently. Nothing in a dashboard warns you.

robots.txt is per bot

OpenAI's crawler page says each setting "is independent of the others". GPTBot governs training use. OAI-SearchBot governs ChatGPT search, and sites that block it "will not be shown in ChatGPT search answers". Perplexity uses PerplexityBot for its index, and Anthropic uses Claude-SearchBot for Claude search. A blanket AI block copied from a template is the most common way to drop out of ChatGPT search by accident.

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

This example allows the search bots and opts out of OpenAI training. That's a legitimate choice. Blocking GPTBot doesn't remove you from ChatGPT search.

The firewall can override robots.txt

OpenAI says eligibility needs both: allow OAI-SearchBot, and confirm your host or CDN accepts traffic from OpenAI's published search bot IP addresses. Bot protection can return a 403 to a crawler your robots.txt allows. Since September 15, 2026, Cloudflare blocks its Training and Agent crawler categories by default on pages that display ads, for new domains and existing free-tier customers, while Search stays allowed. Many small business sites run on default settings, so check rather than assume. Our bot management glossary entry covers how these systems classify crawlers. The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away.

JavaScript hides facts from most AI crawlers

Vercel's crawler study found that ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. Googlebot renders JavaScript, so this mostly affects non-Google engines. Booking widgets, hours loaded from a script, and location finders are common culprits. Put the facts that matter in the HTML the server sends.

Google and Bing indexing

Google says a page must be "indexed and eligible to be shown in Google Search with a snippet" to appear as a supporting link in AI Overviews or AI Mode, with no additional requirements. A stray noindex or nosnippet keeps a page out. Bing's index backs Microsoft Copilot. In Otterly's 2026 experiment, Google deindexed two test sites and Google's AI surfaces stopped citing them, while Copilot kept citing them from Bing. Check each engine's index separately.

Accurate business facts: the second controllable part

AI platforms get specific facts wrong about specific businesses. For a local business the costly errors are concrete: an old address, wrong opening hours, a service you stopped offering, a price from two years ago.

The one positive experimental signal for on-page content is simple: stating a specific fact on a fetchable page gives the engine a sentence it can quote. In practice:

Write your address, phone number, hours and service area as plain text on your site, not only inside an image, map embed or script.
State prices or price ranges where you can, in a sentence, for example "Emergency exams start at $X" with your real figure.
Say what you do and don't offer. "We see emergency patients on Saturdays from 9 to 1" answers the hypothetical question directly.
Make every page agree. If the footer, contact page and about page give different hours, an engine has no way to know which is right.
Correct the same facts on third-party pages you can edit, such as directory listings, so retrieved sources agree with your site.

The local business fact check playbook walks through this list fact by fact. The free checker on this site shows whether AI answers name you for up to three buyer questions, which is a useful snapshot before and after you correct things.

A practical order of work

Work from the gates outward. A fix at a later step can't help if an earlier gate is closed.

Check robots.txt per bot. Confirm OAI-SearchBot, PerplexityBot, Claude-SearchBot and Googlebot are allowed. Decide on training bots separately.
Check the firewall and CDN. Request your homepage and contact page as each AI crawler. Fix any 403 or challenge page.
Check rendering. Load your key pages with JavaScript off. If hours, address, services or prices disappear, move them into server-rendered HTML.
Check index and snippet eligibility. Look in Google Search Console and Bing Webmaster Tools separately. Remove accidental noindex, nosnippet or max-snippet:0 on pages that carry your facts.
Write the facts down plainly. One clear sentence per fact, on a page crawlers can fetch.
Make the facts consistent. Across your own pages first, then on third-party listings you control.
Check what the engines say. Ask the buyer questions your customers ask, engine by engine, and note wrong facts. Repeat after changes, knowing answers vary by query and by whether the engine searched.

For how this fits with the rest of local marketing, see AI local SEO for local businesses.

Common mistakes

Mass-producing city pages. Generating hundreds of near-identical "dentist in [suburb]" pages is the pattern Google's spam policy describes: "Using generative AI tools or other similar tools to generate many pages without adding value for users." In Otterly's experiment, Google removed two AI-generated sites with no manual action warning in Search Console.
Allowing GPTBot to get into ChatGPT. GPTBot governs training only. ChatGPT search uses OAI-SearchBot.
Trusting robots.txt alone. A firewall can still block the crawler.
Adding schema or llms.txt as an AI fix. Google says structured data isn't required for its generative AI features, and Google Search ignores llms.txt. Ahrefs' matched test found AI Overview citations fell 4.6% after pages added schema. Existing markup does no harm, but it isn't a lever.
Hiding key text by accident. A data-nosnippet or nosnippet copied from a template can keep your hours out of AI summaries. See whether hiding text from AI summaries hurts rankings.
Hidden text or instructions aimed at AI crawlers. Cloaking and prompt injection are illegitimate and a reputational risk. Never use them.
Reading one blended score. Engines draw on different indexes and behave differently. Measure each one separately.

Frequently asked questions

Does my Google Business Profile affect AI answers?

The evidence behind this page doesn't include a primary source on how AI answers use business profiles, so we won't claim a mechanism. What Google does document is that AI Overviews and AI Mode cite pages that are indexed and eligible for a snippet, with no extra requirements. Keeping your profile accurate is sound practice for ordinary local search. For AI answers, make sure your own site states the same facts in plain, fetchable text.

Should I create a separate page for every town I serve?

Only where each page adds something real for that area, such as a genuine location, staff, hours or services that differ. Google's spam policy targets many pages generated mainly to manipulate rankings without adding value for users. Thin, templated town pages risk exactly that. One accurate page per real location, plus a clear statement of your service area, is a safer structure than a page for every postcode.

Does schema markup help local businesses appear in AI answers?

There's no credible evidence that it does. Google says structured data isn't required for generative AI search. Ahrefs' matched test of 1,885 pages found no major citation uplift, and a 4.6% drop in AI Overview citations. That study only covered pages already cited heavily, so effects on smaller sites are untested. Existing markup does no harm. Spend your time on crawler access and accurate facts first.

Do AI answers for local queries change depending on where the searcher is?

Often, yes. Google AI Overviews don't appear on every search, and they vary by location and language. ChatGPT may also use saved memories when rewriting a search query, so logged-in results differ from person to person. When you check what engines say about your business, test logged out where you can, note the location and language, and treat each result as a snapshot rather than a fixed position.

Why does an AI assistant give my old address?

Usually one of two reasons. The answer may come from the model's memory, which is frozen at a training cutoff, so a recent move isn't there. Or the engine searched and read a source with the old address, such as an outdated directory listing or an old page on your own site. You can't change the first. You can fix the second by correcting every page and listing you control.

Sources

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.