How to show up in ChatGPT
Showing up in ChatGPT comes down to being the easiest credible source to extract: direct answers near the top of the page, structured data, clean access for AI crawlers, and coverage of the specific questions buyers ask. No one outside the model providers knows the selection criteria precisely - our benchmark of 384 observations found engines agree on citations only 19% of the time - so the reliable approach is empirical: audit the prompts you should win, ship the structural fixes, and re-test the same prompts to see which changes moved the citation.
How does ChatGPT decide which brands to mention?
ChatGPT answers from two different places, and they reward different work. The first is its training: patterns the model absorbed from a large slice of the web up to its cutoff. When you ask it for the best tool in a category and it answers without browsing, it is drawing on what it learned — which means the brands described most consistently across many independent sources tend to surface.
The second is browsing. When ChatGPT searches the live web (in ChatGPT search and connected modes), it fetches current pages and can quote and cite them directly. This path rewards fresh, crawlable, clearly written pages — and it is the path you can influence fastest, because you do not have to wait for a future training run.
Why isn't my brand showing up in ChatGPT?
Usually one of three reasons. Your pages may be unreadable to the model — blocked crawlers, content that only appears after JavaScript runs, or answers buried far down the page. Your brand may be thinly described, present only on your own homepage rather than corroborated across reviews and editorial coverage. Or competitors may simply own the third-party sources the engine leans on for your category.
Step 1 — Let ChatGPT's crawlers read your site
If an engine cannot fetch a page, it can never cite it. Make sure your robots.txt and CDN do not block OpenAI's crawlers, and confirm your key pages render their content in raw HTML rather than only after client-side scripts execute.
- Allow GPTBot, which gathers content that can inform training.
- Allow OAI-SearchBot, which powers ChatGPT's live search and citations.
- Verify the page returns full content to a plain fetch, not just to a browser.
- Remove login or paywall gates from anything you want cited.
Step 2 — Structure pages so ChatGPT can extract them
Lead with a direct, self-contained answer to the page's core question in the first paragraph, then expand. Use headings phrased the way people ask questions, keep paragraphs short, and add Article and FAQPage structured data so the page is easy to parse and attribute. A page that answers one question cleanly is far more quotable than a long, unfocused one.
Step 3 — Earn the third-party sources ChatGPT trusts
Most AI citations point to pages you do not own — reviews, roundups, comparison articles, and community threads. When ChatGPT browses your category, being named in those sources is often the single biggest lever, because the model treats independent corroboration as stronger evidence than your own marketing copy.
Step 4 — Track the prompts and prove the change
Pick the category, comparison, and alternative prompts a buyer would actually type, run them in ChatGPT, and record whether you are named and cited and who appears instead. Ship one fix at a time, then re-run the same prompts. Because the same prompt can return slightly different answers on different runs, repeated testing is what separates a real gain from noise.
Frequently asked questions
Not for organic citations. You earn them through crawlability, extractable content, broad corroboration across trusted sources, and freshness — not payment.
When it browses the live web it shows citations you can click. When it answers purely from training it usually does not, which is why broad, consistent coverage across the web matters as much as any single page.
Browsing-based citations can reflect a fresh, crawlable page within days. Training-based recall changes far more slowly, on the timescale of model updates, so corroboration is a long game.
GPTBot collects content that can inform model training; OAI-SearchBot fetches pages to answer live queries with citations. Blocking either reduces a different path to being mentioned.
Typically because the competitor is described more clearly on extractable pages and named more often across the third-party sources ChatGPT trusts for your category. Closing that source gap is the fix.