Why doesn't AI mention my business?
Because the engine can't read, verify, or trust your facts yet — almost never because you're too small, too new, or not good enough. AI answers are built from sources that pass those three checks, and most business websites fail at least one silently. The failing check is findable in minutes, and most of the fixes are free.
Published 2026-07-18 · Updated 2026-07-18 · Grounded AEO
When ChatGPT, Perplexity, or Google's AI answers a question your business should win and leaves you out, one of five things is true — listed here from most to least common:
→ Your facts disagree with themselves. Your site says one address, an old directory says another, a social profile says a third. Engines don't pick a winner — they conclude you can't be trusted and move on.
→ Crawlers can't read you. Your robots.txt blocks AI crawlers (often from a rule added years ago), or your content only exists in JavaScript the crawler never runs.
→ Nothing corroborates you. The engine found your site but nothing else agrees — no consistent listings, no third-party mentions, nothing to cross-check against.
→ Your facts aren't machine-readable. A human can tell what you do; a parser extracting name, place, and offering from your prose can't.
→ Nobody in your niche is citable yet. The engine answers generically because it has no one it can safely name. That's not a wall — it's an open door for whoever moves first.
Match what you're seeing to the row that fits. Each symptom points at a different failure:
AI names competitors, never you
CAUSETheir facts pass the read/verify/trust checks; yours fail at least one
FIXRun the checks below in order — the failing one is different for every business
AI describes you but gets facts wrong — old address, dead phone, wrong hours
CAUSEThe engine is answering from stale scraps: forgotten directories, an old profile
FIXFix the drift at its source, then publish current facts somewhere you control
AI says it has no information about you
CAUSECrawlers cannot read your site — blocked in robots.txt, or the content only exists in JavaScript
FIXCheck robots.txt for AI-crawler blocks; make your basics server-rendered
AI answers the question generically and names nobody
CAUSENobody in your niche passes the checks yet — the engine is working from the old phone book
FIXBest case on this page: the first business to publish verifiable facts becomes the answer by default
You rank fine in Google Search but never appear in AI answers
CAUSERanking and citation are different tests — keywords win lists, verifiable facts win answers
FIXAdd structured data and an llms.txt; consistency matters more than keywords here
Everything below works without buying anything from anyone, including us. In order:
1. Open yourdomain.com/robots.txt. Look for lines blocking GPTBot, ClaudeBot, PerplexityBot, Google-Extended, or a blanket Disallow: /. If the crawler can't get in, nothing else on this list matters.
2. Ask the engines about yourself. Literally: “What do you know about [your business] in [your town]?” in ChatGPT and Perplexity. Write down every wrong or missing fact. Each one traces back to a stale source you can now hunt down.
3. Kill the drift. Pick the one exact name, address, and phone number, then make it identical everywhere it appears: your site footer, Google Business Profile, Apple and Bing listings, the directories from step 2. Boring, unglamorous, load-bearing.
4. Add structured data. A schema.org LocalBusiness or Organization JSON-LD block on your homepage turns your facts from prose an engine must guess at into fields it can extract.
5. Write an llms.txt. A plain-text guide to your business that AI crawlers fetch first. Our llms.txt guide has a complete worked example — about twenty minutes of work.
6. Wait a cycle, then re-measure. Crawlers revisit on cycles measured in weeks. Re-ask the same questions monthly and log the answers. That log is your progress bar.
What a service like ours adds on top is the part you can't easily do alone: verification (proving the facts are yours via your domain), a signed machine-readable record engines and agents can pull directly, and monitoring that re-asks the questions for you. But the six steps above are the foundation either way — and they're yours free.
Answer engine
Any AI that answers questions directly instead of returning links — ChatGPT, Perplexity, Google AI Overviews, Claude, Copilot.
Citation
Being named or quoted inside the answer itself. The AEO equivalent of ranking #1 — except there are usually only one or two slots.
AEO
Answer Engine Optimization: making your business readable, verifiable, and trustworthy to answer engines. Sibling of SEO, different test.
Grounding
What an engine does when it bases its answer on retrieved, checkable sources instead of its training memory. Grounded answers cite; ungrounded answers guess.
llms.txt
A plain-text guide to your business at yourdomain.com/llms.txt that AI crawlers fetch first. Twenty minutes to write.
Structured data
Machine-readable facts (schema.org JSON-LD) embedded in your pages so engines can extract who you are without guessing from prose.
NAP drift
Name/address/phone inconsistency across the web — the single most common reason engines refuse to trust a local business.
v=AEO1 record
A DNS TXT record at _aeo.yourdomain.com proving a verified fact catalog exists for your domain. Open standard: aeorecord.org.
Hallucination
When an engine states something false with confidence. For businesses, usually not random — it is an echo of a stale source somewhere.
Every check above gets stronger with time. Engines don't just ask “are these facts consistent?” — they ask “have these facts stayed consistent?” A record that has been published, refreshed, and re-verified across months outweighs an identical record created yesterday, the same way domain age became a quiet ranking signal in classic search.
This cuts both ways. Drift also compounds: every year a forgotten listing sits wrong, it gets re-scraped into more places the engine reads. The cheapest time to fix your facts was when they first drifted. The second cheapest is now.
And because most niches still have nobody who passes the checks, tenure is the moat that's still unclaimed almost everywhere. Whoever starts the clock first keeps that lead by default.
Can I pay ChatGPT or Google to include my business in AI answers?+
No. There is no slot to buy inside the organic answer itself. Ads exist around some AI surfaces, but the answer is assembled from sources the engine can read, verify, and trust. The only durable lever is becoming one of those sources.
How long until AI starts mentioning me?+
Nobody can honestly promise a date. Crawlers revisit on cycles measured in weeks, and each engine runs its own. What you can do is measure: ask the engines your target questions once a month and log whether you appear. Movement shows up in that log before it feels like traffic.
AI is saying something wrong about my business. How do I fix it?+
Trace the wrong fact to its source — it is almost always a stale directory listing, an old profile, or a page you forgot exists. Fix it there, publish the current fact somewhere you control (your own site, in structured form), and wait a crawl cycle. Engines drop a stale source when a fresher, more consistent one exists.
Do I need to rebuild my website?+
No. Every fix on this page is additive: a robots.txt check, a block of structured data, a plain-text llms.txt file, consistency edits. Your design does not need to change — engines read structure, not aesthetics.
Is this just SEO with a new name?+
They overlap but pass different tests. SEO is about ranking in a list of links; AEO is about being quoted inside the answer. Ranking rewards keywords and links. Citation rewards facts an engine can verify. You can rank #1 and still never be mentioned by AI.
How do I verify my business information for AI?+
Three layers, in order of strength. First, self-declared: publish your facts in structured data and llms.txt on your own domain — that puts them in your words, on your property. Second, corroborated: make every major listing agree with them. Third, verified: publish a record that proves domain ownership, like a v=AEO1 DNS record pointing at a maintained fact catalog — the strongest signal, because it is machine-checkable and cannot be faked by a third party.
How do I control what AI says about my company?+
You cannot edit an engine's answer directly, and no one honest will sell you that. What you control is the supply: the sources engines pull from. Publish current, structured, consistent facts on your own domain, clean up the stale sources echoing old information, and keep a verified record engines can check. Engines follow the best available source — so become it.
What is a v=AEO1 record?+
A DNS TXT record at _aeo.yourdomain.com that tells AI engines, in one machine-checkable line, that a verified fact catalog for your business exists and where to find it. It is an open standard — the spec lives at aeorecord.org and anyone can publish one, with or without us.
My business is brand new. Should I wait until this matters more?+
The opposite. Trust signals compound with tenure — a record that has been published, maintained, and refreshed for two years beats an identical one created yesterday. The month you start is the month the clock starts.
Which check do you fail?
The free audit runs them all against your domain and shows every gap in plain English. 30 seconds, no account.