The short answer
ChatGPT doesn’t mention your business because it can’t confirm you exist, what exactly you do, or that any independent source vouches for you. AI assistants don’t rank businesses the way Google ranks pages. They assemble an answer from whatever they can retrieve and cross-check in seconds: crawlable pages on your site, your listings in directories, reviews, and third-party mentions. If those sources are thin, contradictory, or blocked to the crawlers these assistants use, you were never a candidate. You weren’t rejected. You were invisible.
The good news is that almost everything in that list is verifiable and fixable. The bad news is that none of it can be bought, rushed, or guaranteed.
Why this suddenly matters
BrightLocal’s Local Consumer Review Survey, published in March 2026 from a sample of 1,002 US adults, found that 45% of consumers now use AI tools to get local business recommendations, up from 6% the year before. That puts AI behind only Google and Facebook as a discovery channel, ahead of Yelp and TripAdvisor. The same study found 88% of AI users fact-check what the assistant tells them.
That last number is the useful one. People aren’t blindly buying from whoever ChatGPT names. They’re using the assistant to build a shortlist, then verifying. So the goal isn’t to “win” the answer. It’s to be on the list, and to survive the verification that follows.
What actually happens when someone asks
Three different pipelines are at work, and they don’t share plumbing.
ChatGPT runs a live retrieval step. OpenAI’s own help documentation states that ChatGPT search uses third-party search providers, including Microsoft Bing, plus content from licensed partners. OpenAI’s developer documentation lists a specific crawler, OAI-SearchBot, for surfacing sites in ChatGPT search, and says plainly that sites opted out of it will not be shown in ChatGPT search answers. It also notes that robots.txt changes take roughly 24 hours to register.
Google’s AI Overviews and AI Mode use the regular Google index. Google’s own AI optimization guide says a page only needs to be indexed and eligible to appear with a snippet to be eligible as a supporting link. No separate submission, no separate crawler.
Perplexity runs its own crawler and cites sources inline, which makes it the easiest platform to audit — you can see exactly which page it pulled from.
| Assistant | Where it gets web results | What you control directly |
|---|---|---|
| ChatGPT search | Third-party providers including Bing, plus licensed partner content | Allow OAI-SearchBot in robots.txt; claim Bing Places; verify in Bing Webmaster Tools |
| Google AI Overviews / AI Mode | The standard Google index via Googlebot | Normal indexing and Google Business Profile; monitor the Search Console generative AI report |
| Perplexity | Its own crawler plus on-demand page fetches | Allow PerplexityBot; make sure cited pages answer the question well |
| Gemini / Copilot | Google and Bing indexes respectively | Same underlying work as the two above |
The practical takeaway: there is no single “AI listing” to submit. You are optimizing two indexes, a handful of directories, and your own site.
The five reasons you’re not being named
1. A crawler you’ve never heard of is blocked
Plenty of small business sites were built with a blanket robots.txt rule, a security plugin, or a Cloudflare bot-fight setting that blocks AI user agents. If OAI-SearchBot is disallowed, OpenAI’s documentation is explicit that you won’t appear in ChatGPT search answers.
Do this: open yourdomain.com/robots.txt in a browser. Look for anything blocking OAI-SearchBot, PerplexityBot, or Googlebot. Then check your server logs or CDN dashboard for those user agents over the last 30 days. Zero visits from a crawler means it never got in. Note that GPTBot is a separate agent used for model training — you can allow the search crawler and block the training one if you prefer. They’re independent settings.
2. Your business facts disagree with each other
Assistants build confidence through agreement. If your suite number is on your website but missing from Yelp, your phone number changed two years ago but three directories still list the old one, and your hours differ across four platforms, the assistant has no confident fact to state. It picks the competitor whose details are boring and identical everywhere.
Do this: write your name, address, phone, and hours in one document exactly as they should appear — including whether you use “Suite 16” or “Ste 16.” Then correct them everywhere: Google Business Profile, Bing Places, Apple Business Connect, Yelp, Facebook, your chamber of commerce, industry associations, and your own site footer. Budget three to four hours. It is tedious and it is the single highest-leverage thing on this list.
3. You’re absent from the sources being read
Assistants lean heavily on aggregators and editorial sources because those cover a whole category at once. A competitor gets named because they appear in a “best in McKinney” roundup, a chamber directory, a trade association member list, and a local news mention. You have a website and nothing else.
Do this: pick five genuine third-party surfaces you can legitimately be listed on this quarter. Local chamber, one or two industry associations, a regional business publication, a supplier’s partner page, a local sponsorship. These are real relationships, not link buying. Directory and editorial mentions are the raw material assistants use to say “several sources name this company.”
4. Your pages don’t answer the question being asked
People type full sentences into assistants: “who does commercial roof repair near McKinney and handles insurance claims?” If your site has a services page listing “Roofing” with a photo grid, there is nothing to retrieve. The competitor with a page headed “Do you handle insurance claims for storm damage?” followed by a straight two-paragraph answer is trivially quotable.
Do this: list the ten questions you answer on the phone every week. Give each one a real heading and a direct answer in the first two sentences, before any setup. Include prices or ranges if you can, service area names, turnaround times, and what you don’t do. Specificity is what makes a page retrievable. This is ordinary clear, question-led content writing — not a new discipline.
5. Your reviews are thin, old, or unanswered
Reviews are the closest thing to independent verification these systems have. A profile with 11 reviews, none in the last 18 months, gives an assistant nothing to summarize. Volume, recency, and the actual words matter — reviews that mention the specific service and the specific city give an assistant language to match against a query.
Do this: build a review request into the moment a job closes. Ask by name, ask on the day, and never script the wording. Respond to every review, including the good ones. Aim for steady flow rather than a burst.
A realistic 30-day sequence
- Week 1 — audit. Check robots.txt and crawler logs. Verify your site in Bing Webmaster Tools (free, about ten minutes) and confirm Google Search Console is connected. Ask each assistant ten real buyer questions in your category and city, and write down exactly who gets named and which sources are cited.
- Week 2 — consistency. Fix your business facts everywhere. Claim Bing Places if you never have; it’s free and it feeds the Microsoft ecosystem that ChatGPT search draws on.
- Week 3 — content. Publish or rewrite five pages that answer your five most common buyer questions in plain language, with your service area named explicitly.
- Week 4 — proof. Start the review habit, and pursue two or three genuine third-party listings or mentions.
Then wait. Retrieval indexes update on their own schedule, and model behavior shifts without notice. Re-run your ten test prompts monthly and log the results in a spreadsheet. That log is your only real measurement.
What you can also measure
Google rolled its Search Console generative AI performance report out to all sites worldwide on August 31, 2026. It shows impressions from AI Overviews and AI Mode broken out by page, country, and date. It doesn’t yet include query-level data, but it’s the first free, first-party window into AI visibility, and it’s worth checking monthly. Bing Webmaster Tools gives you the equivalent view on the Microsoft side.
Honest limits — what doesn’t work
You can’t buy the recommendation. ChatGPT does carry advertising now, and sponsored cards appear beneath answers on free tiers. But OpenAI states that ads run on separate systems from the chat model and that advertisers cannot shape, rank, or alter ChatGPT’s responses. Buying an ad slot is a legitimate media purchase. It is not a shortcut to being named in the answer itself.
Skip the magic files. An llms.txt file has been widely promoted as an AI-era robots.txt. Google’s Search Relations team, including Gary Illyes and John Mueller, has said publicly that Google doesn’t support it and doesn’t plan to, comparing it to the long-dead keywords meta tag. No major model provider has committed to using it as a ranking signal. Adding one is harmless; expecting results from it is not.
Google’s own position is worth reading. Its AI optimization guide states there’s no special schema markup required for AI features, no need to chunk content into fragments, and that “optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” It also explicitly cautions against third-party AEO and GEO services promising more. Structured data is still worth adding — it earns rich results and removes ambiguity about your hours and location — but treat anyone selling it as a secret AI hack with suspicion.
Nobody can guarantee a mention. These systems are probabilistic. The same question asked twice can produce different names. Any agency promising guaranteed AI placement is either misinformed or lying.
Where to start
Run the ten-prompt test this week. If you’re not named and a competitor is, work down the five causes above in order — crawler access, factual consistency, third-party presence, question-led pages, reviews. Most businesses find the problem in the first two.
If you’d rather not run the audit yourself, that groundwork is exactly what our SEO and SEM work covers: crawler and index diagnostics, listing cleanup across the platforms that feed these assistants, and monthly prompt tracking so you can see movement. The content side handles the question-led pages, and our AI solutions team can set up automated monitoring so you’re not checking prompts by hand.
Either way, the work is the same, and it’s the kind of work that holds up regardless of which assistant your customers use next year.
Frequently asked questions
Can I pay to get my business recommended by ChatGPT?
No. ChatGPT does show sponsored cards beneath answers on free tiers, but OpenAI states that ads run on separate systems from the chat model and that advertisers cannot shape, rank, or alter its responses. You can buy an ad placement. You cannot buy the organic recommendation itself, and no agency can guarantee one.
How long does it take to start appearing in AI assistant answers?
Expect months, not weeks. Robots.txt changes register with OpenAI’s systems in about 24 hours, but index refreshes, new directory listings, and review accumulation all move on their own schedules. Run the same ten test prompts monthly and log who gets named. Steady movement over a quarter is a realistic expectation.
Do I need special schema markup or an llms.txt file for AI search?
No. Google’s AI optimization guide states there is no special schema markup required for AI features and no need to chunk content into fragments. Google’s Search Relations team has said publicly it does not support llms.txt and does not plan to. Standard structured data is still worth having for rich results and factual clarity.
Does my Google Business Profile affect what ChatGPT says?
Only indirectly. ChatGPT search draws on third-party providers including Bing, so Google Business Profile data reaches it secondhand at best. Claim your free Bing Places listing as well, and make sure your name, address, phone, and hours match exactly across both plus Yelp, Apple, and Facebook.
How do I check whether AI crawlers can actually reach my site?
Open yourdomain.com/robots.txt in a browser and look for rules blocking OAI-SearchBot, PerplexityBot, or Googlebot. Then check your server logs or CDN dashboard for those user agents over the last 30 days. Zero recorded visits means the crawler never got in, which alone can explain your absence.
