On this page
AI assistants like ChatGPT and Perplexity find local businesses by searching the web and reading pages, then citing some of them. To be one of those sources, let their search crawlers (such as OAI-SearchBot and PerplexityBot) read your site, publish clear service and area pages in plain text, keep listings consistent, and earn real reviews. No method guarantees a mention.
Key takeaways
- AI assistants answer local questions by searching the web and citing pages, so your site has to be readable by their crawlers.
- OpenAI uses OAI-SearchBot for ChatGPT search; blocking it in robots.txt keeps you out of ChatGPT search answers.
- GPTBot (training) and OAI-SearchBot (search) are separate, so you can allow one and block the other.
- Google says its AI features need nothing beyond normal search eligibility.
- Nobody can guarantee an AI mention, and anyone who does is guessing.
How do AI assistants find local businesses?
When someone asks ChatGPT "who does gutter cleaning near me" or asks Perplexity for a recommended electrician, the assistant often doesn't answer from memory alone. It can run a web search, read a handful of pages, and write a summary with links to the sources it used. The businesses it names come from what those sources say.
That means the question "how do I show up in ChatGPT" mostly breaks down into two simpler ones. Can the assistant's crawler read your pages? And when it reads them, do your pages, listings and reviews make it obvious which services you offer, which towns you cover, and that real customers have hired you?
Each company runs its own crawlers and search systems, and they don't publish how they choose which businesses to name. What they do publish is how their crawlers work and how to let them in. That's where to start.
Which crawlers should you allow?
Your site's robots.txt file tells crawlers what they may read. Security plugins, hosts and firewall services can block AI crawlers, sometimes without the owner knowing. Cloudflare, for example, announced in July 2025 that it would block AI crawlers by default and ask every new domain whether to allow them. Check yours.
| Crawler | Company | What it's for, per the company |
|---|---|---|
| OAI-SearchBot | OpenAI | Surfacing websites in ChatGPT's search features |
| GPTBot | OpenAI | Crawling content that may be used to train its models |
| ChatGPT-User | OpenAI | Visiting a page when a user's request calls for it |
| PerplexityBot | Perplexity | Surfacing and linking websites in Perplexity search results |
| Perplexity-User | Perplexity | Visiting a page to answer a user's question |
| Googlebot | Google Search, which also feeds AI Overviews and AI Mode | |
| Google-Extended | A robots.txt token for opting out of Gemini model training, with no effect on Search |
OpenAI's crawler documentation says OAI-SearchBot is used to surface websites in search results in ChatGPT's search features, and that sites which block it won't appear in ChatGPT search answers, though they may still show up as plain navigational links. It also notes that changes to robots.txt can take about 24 hours to take effect on its side. ChatGPT-User is different: OpenAI says it visits pages when a person asks, isn't used to decide what can appear in search, and, because a user started the action, robots.txt rules may not apply to it.
The useful detail for owners is that training and search are separate. If you don't want your content used to train models, you can block GPTBot and still allow OAI-SearchBot. Perplexity works the same way: its documentation says PerplexityBot surfaces and links websites in Perplexity search results and isn't used to crawl content for training foundation models, and it recommends allowing PerplexityBot if you want to appear. The same page says Perplexity-User, which fetches a page when someone asks about it, generally ignores robots.txt rules because a user requested the fetch.
Open yourdomain.com/robots.txt in a browser. If you see Disallow: / under OAI-SearchBot, PerplexityBot or a wildcard user agent, the assistants that respect it can't read your site. Also check whether a firewall or bot-protection setting on your host is blocking them.
What about Google's AI Overviews?
Google is unusually direct about this. Its Search Central documentation says there are no additional requirements to appear in AI Overviews or AI Mode, and that you don't need to create new machine-readable files, AI text files or markup. To be eligible, a page must be indexed and eligible to be shown in Google Search with a snippet.
Google's AI training control is separate. Google's crawler documentation describes Google-Extended as a robots.txt token for managing whether your content may be used to train future Gemini models, and says it does not affect a site's inclusion in Google Search and isn't a ranking signal. Blocking Googlebot is what would keep you out of Search and its AI features.
In other words, if your page is a good result for normal Google Search, it's eligible for the AI features. If it isn't indexed, nothing else matters. Google Search Console will tell you which of your pages are indexed and why any aren't.
What helps an AI assistant recommend you?
None of these companies publish how they pick which businesses to name. The safest bets are the things that make a business easy to describe accurately, and they help with ordinary search too.
- A page for each service, in plain text. If your services are listed only inside an image or a PDF, a crawler may miss them. Write "We replace asphalt shingle roofs in these towns" as text on a page about roof replacement.
- Clear service-area information. Name the towns and counties you work in. Assistants answering "near me" questions need a location to match.
- Consistent listings. Your business name, phone and address (or service area) should match across your site, your Google Business Profile, Bing Places, Apple Business (which Apple says replaced Apple Business Connect in 2026) and trade directories. Conflicting details make you harder to describe with confidence.
- Real reviews in public places. Review platforms are public pages an assistant can read and cite. You can't control what they quote, but you can ask every customer for an honest review.
- Structured data that matches the page. Google's LocalBusiness structured data documentation says the markup can tell Google about details like business hours and departments. It gives machines facts in a format they don't have to guess at, though Google is clear it isn't required for AI features. Keep it identical to what the page shows.
- Specific, factual answers. "We service tankless water heaters from these brands" is something an assistant can quote. "We go above and beyond" isn't.
- Being mentioned elsewhere. Local news, supplier dealer pages, trade associations. An assistant that reads several sources has more chances to find you when other sites describe your work, not only your own.
What doesn't work, or isn't proven
Be skeptical of anyone selling a shortcut. OpenAI publishes no ranking formula for which businesses ChatGPT names. The paid product its crawler documentation does mention is advertising: OAI-AdsBot checks pages submitted as ads on ChatGPT. Buying an ad is not the same as being named in an answer. Stuffing pages with phrases like "best plumber in town" doesn't make an assistant believe it.
Fake reviews are a legal problem as well as a policy one. In August 2024 the FTC announced a final rule (now 16 CFR Part 465) that bans fake consumer reviews and offering payment or other incentives on the condition that a review express a particular sentiment, positive or negative, and gives the FTC the power to seek civil penalties from anyone who violates it knowingly. That is general information, not legal advice; ask a lawyer before you offer anything in exchange for a review.
Files like llms.txt are a low-cost addition, but Google says AI text files aren't needed for its AI features, and neither OpenAI's nor Perplexity's crawler documentation says the file affects which businesses get named (both companies do publish an llms.txt for their own developer docs). We explain the format and its limits in what llms.txt is.
And there's no guarantee. The same question asked twice can produce different businesses. Assistants also change how they search and cite over time. The work above improves your odds across all of them; it doesn't lock in a result with any of them.
A step-by-step checklist
If you want to work through this in an afternoon, here is the order we'd use.
- Read your robots.txt. Confirm that OAI-SearchBot, PerplexityBot and Googlebot aren't blocked. Decide separately whether to allow GPTBot.
- Check your host's bot settings. Some hosting and firewall services block unfamiliar crawlers at the network level, which robots.txt can't override. Look for a bot management or AI crawler setting.
- Confirm indexing in Google Search Console. Every service page you care about should be indexed. If one isn't, find out why before writing anything new.
- Read each service page as a stranger. Could someone who has never heard of you say what you do, where, and roughly how it works after one read? If not, rewrite it in plain text.
- Line up your listings. Make your name, phone and service area identical on your site, Google Business Profile, Bing Places, Apple Business and the two or three directories that matter in your trade.
- Add or fix LocalBusiness markup. Match it exactly to the visible page.
- Start a review routine. Ask every customer after every job.
- Set a monthly reminder to run your test questions in a few assistants and note what changed.
Most of this is a one-time cleanup followed by small, regular upkeep that the office can own. It also helps your regular Google rankings, so none of the effort is wasted if AI search changes again.
How to tell if it's working
Measurement is still rough, but there are signals you can watch:
- Referral traffic in analytics. Visits from chatgpt.com, perplexity.ai and similar domains can show up as referrals, though some arrive with no referrer and look like direct traffic.
- Search Console. Per Google's AI features page, clicks that come through AI Overviews and AI Mode are counted in the Performance report under the Web search type, as part of your overall search traffic.
- Ask your callers. Add "AI assistant" as a lead-source option in your CRM and have the office ask how people found you.
- Run the searches yourself. Once a month, ask a few assistants the questions your customers ask, in your area, and note who gets named and which pages are cited. Those cited pages tell you what kind of source the assistant trusts for your trade. Our free AI visibility check writes eight of those questions for your trade and town and keeps the tally in your browser.
This is part of the search work we do on our SEO and AI search service, and it builds on the same foundation as local SEO for contractors. If you want to know whether your site is readable by these crawlers right now, our free system quote checks it.
Frequently asked questions
Can I pay ChatGPT to recommend my business?
OpenAI does sell ads on ChatGPT (its crawler documentation describes OAI-AdsBot, which checks pages submitted as ads), but an ad is not the same as being named in an answer, and OpenAI publishes no paid route into the answers themselves. To be eligible for ChatGPT search answers, OAI-SearchBot has to be allowed to read your site.
Should I block GPTBot?
That's a business decision about training. Blocking GPTBot doesn't block OAI-SearchBot, which is the crawler OpenAI uses for ChatGPT search results.
Does Google need special files for AI Overviews?
No. Google says no new machine-readable files, AI text files or markup are needed. A page must be indexed and eligible to show with a snippet.
Why does ChatGPT name a competitor instead of me?
Usually because the sources it read describe that business more clearly, more consistently, or more often. Look at which pages it cites and compare them to yours.
Sources
- OpenAI: Overview of OpenAI crawlers
- Perplexity: Perplexity crawlers
- Google Search Central: AI features and your website (updated 2025)
- Google: Google's common crawlers (Google-Extended), 2026
- Google Search Central: Local business structured data, 2026
- Cloudflare press release, July 1, 2025
- Apple Newsroom: Introducing Apple Business, March 2026
- FTC: Final rule banning fake reviews and testimonials, August 2024
General information for service-business owners, not legal, tax or financial advice. Rules vary by state; check with a qualified professional for your situation.