Guide

How do I get my business to show up in ChatGPT?

Showing up in ChatGPT comes down to six things: get into Bing's index, allow OpenAI's OAI-SearchBot crawler, serve your content as real HTML that reads without JavaScript, give each page a self-contained answer to one specific question, add LocalBusiness and FAQPage structured data, and keep your Google Business Profile complete and consistent.

The part that surprises most owners: ChatGPT's search feature does not read Google's index. Ranking well on Google does not put you in ChatGPT.

Why your rankings don't transfer

AI assistants read a different map

Google spent twenty years teaching everyone that search visibility is one thing you either have or don't. AI assistants broke that assumption. ChatGPT's search feature leans on Bing's index plus OpenAI's own crawler. Google's AI Overviews read Google. Perplexity and Claude run their own crawlers. These are separate pipelines with separate requirements, and being excellent in one says almost nothing about the others.

The second difference matters more. Traditional search returns a ranked list of pages, so the page is the unit. AI assistants assemble an answer out of passages pulled from several sources, so the passage is the unit. A page that covers eight topics competently will lose to a page that answers one of them completely, because only the second one produces a clean chunk worth quoting.

The six things

What actually decides whether you appear

In dependency order — the later ones can't compensate for skipping the earlier ones.

  1. Get into Bing's index

    Do this first

    ChatGPT's search feature leans heavily on Bing. If Bing hasn't indexed your site, ChatGPT has very little to cite, no matter how well you do on Google. Add the site to Bing Webmaster Tools and submit your sitemap — you can import the whole property from Google Search Console in a couple of clicks. This is the cheapest item on the list and the one most often skipped.

  2. Let the AI crawlers in

    robots.txt

    Check robots.txt and name the crawlers explicitly rather than relying on a wildcard. OpenAI runs two that matter and they do different jobs: OAI-SearchBot powers ChatGPT search, and GPTBot collects training data. Blocking OAI-SearchBot removes you from ChatGPT's search answers entirely. Blocking only GPTBot keeps you visible in search while opting out of model training — a reasonable position, and a deliberate one. OpenAI notes robots.txt changes can take about a day to take effect.

  3. Serve real HTML

    Most common failure

    If your content only appears after JavaScript runs, a crawler may see an empty shell. This is the most common silent failure on modern small business sites, because the page looks perfect in your browser and blank to everything else. Test it honestly: fetch your own page with curl, or disable JavaScript and reload. If the text isn't there, nothing else on this list can help you.

  4. One page, one question, answer first

    Content shape

    Because engines retrieve passages, the first paragraph has to answer the question completely and stand on its own with no surrounding context. Put the number, the price or the direct yes-or-no in the opening sentences, then explain underneath. A page titled after the exact question a customer would type, answering only that question, beats a services page that mentions it in passing.

  5. Add structured data

    schema.org

    Structured data tells a machine what your page means instead of making it guess. For a local business the load-bearing pieces are LocalBusiness or Organization with your real address and service area, FAQPage for question-and-answer blocks, and a sameAs list linking every profile you control — Google Business Profile, LinkedIn, Facebook, industry directories. That list is what lets a model see one corroborated business rather than several scattered records.

  6. Complete your Google Business Profile

    Local queries

    For anything location-flavoured, your Google Business Profile does more work than your homepage. Assistants pull business name, categories, service area, hours and review content from it. Keep the name, address and phone number identical everywhere they appear — a stale phone number in an old directory is noise that reduces a model's confidence in recommending you at all.

What it looks like done right

You can verify all of this on this page

Every technique above is running on the page you're reading, and none of it is hidden. Fetch this URL with curl and you'll get the complete article text as static HTML, because all six marketing pages here are prerendered at build time rather than assembled in your browser. View the source and you'll find Organization, Service, FAQPage and BreadcrumbList structured data. Read robots.txt and you'll find OAI-SearchBot, GPTBot, PerplexityBot, ClaudeBot, Applebot and the rest named explicitly and allowed, with the SEO scraper bots blocked.

We mention this because it's checkable, which is the only kind of credential worth anything in this field. If someone offers to make your site AI-visible, fetch their own site with JavaScript disabled first.

Measurement

How to tell whether it's working

Citation isn't a rank you can look up. It has to be sampled.

Ask the question yourself

Open an assistant with live search and ask what a customer would ask — not your business name. See whether you're named. Repeat it a few times and on different phrasings, because the result is probabilistic and varies by session.

Watch your server logs

Look for OAI-SearchBot, PerplexityBot and ClaudeBot in your access logs. Crawler visits are the leading indicator: they tell you you're eligible to be cited well before any citation appears.

Watch your analytics

Referrals from chatgpt.com, perplexity.ai and copilot.microsoft.com are the clearest proof that an AI answer sent you a real person. Set this up before you start, or you'll have no baseline to compare against.

Skip these

Four things that don't move the needle

Treating llms.txt as the answer

llms.txt is a proposed convention for pointing assistants at your important pages. No major engine has committed to using it as a ranking input, and we monitor sites that get cited constantly without one. Worth adding — it's a ten-minute job and costs nothing. Not worth believing in as a strategy.

Keyword stuffing, reheated

Repeating “best plumber Salt Lake City” does nothing for a system that reads for meaning rather than matching strings. It reads as noise, and noise is what gets a source skipped.

Publishing AI-written pages at volume

Thin pages that restate common knowledge give a model nothing it can't already produce. What earns a citation is something only you can say: your real prices, your actual process, the specific thing you learned doing the work.

Blocking every crawler to be safe

A blanket disallow does prevent training use — and it also makes you uncitable. Decide the two questions separately, because the crawlers are separate: training access and search visibility are not the same choice.

FAQ

Common questions about AI search

Does ChatGPT use Google's results?

No. ChatGPT's search feature draws on Bing's index together with OpenAI's own crawler, OAI-SearchBot. This is why a business can rank on page one of Google and be completely absent from ChatGPT: nothing about a Google ranking puts you in Bing's index or in front of OpenAI's crawler.

Will blocking GPTBot hurt my visibility in ChatGPT?

No, as long as you leave OAI-SearchBot allowed. They are separate crawlers with separate jobs: GPTBot gathers training data, OAI-SearchBot powers search. Blocking GPTBot alone opts you out of model training while keeping you eligible to appear in ChatGPT's search answers. Blocking OAI-SearchBot is the one that makes you disappear.

Do I need an llms.txt file?

It's cheap and harmless, so add one — but don't expect it to do the work. No major engine has committed to using llms.txt as an input, and plenty of frequently-cited sites don't have one at all. Real HTML, a direct answer in the first paragraph, structured data and presence in Bing's index are what actually determine whether you get cited.

How long does it take to start showing up?

Crawler permission changes are fast — OpenAI says robots.txt updates take effect in roughly a day. Everything after that is slower: the page has to be crawled, indexed, and then judged the best available answer to a particular question. Expect weeks rather than days, and measure crawler visits and search impressions in the meantime, because those move first.

Is this just SEO with a new name?

The discipline is the same and much of the technical work overlaps — crawlability, structured data, fast clean HTML. What changes is the target. SEO optimises a page to rank in a list. AI search optimises a passage to be quoted as an answer, which rewards being specific, self-contained and genuinely informative in ways that ranking never strictly required.

Can I pay to appear in AI answers?

Not in the cited sources themselves. Citations are earned by being the most useful retrievable answer, which is the good news for small businesses: a focused page that answers one real question can be cited alongside companies spending far more than you.

Ready to level up your business?

Tell us what you need — we'll come back with a price and a timeline.

Chat on WhatsAppStart the conversation