How to get cited by ChatGPT.

Key takeaways

• Check crawler access first. Some sites block OAI-SearchBot at the CDN without knowing it, and nothing else matters until that is fixed.
• Third-party sources do most of the work. ChatGPT weights what it did not get from you.
• Write passages that survive being quoted alone.
• Your own site is necessary and not sufficient.
• Measure monthly. Answers move, and one reading is a snapshot rather than a signal.

Step 1: make sure ChatGPT can reach you

This sounds trivial and it is the single most common silent failure. OpenAI uses distinct crawlers: OAI-SearchBot for its search index, ChatGPT-User for live fetches when someone asks about you, and GPTBot for training. Blocking the first two removes you from answers entirely.

CDNs increasingly block AI crawlers by default, and that block overrides whatever your robots.txt says, because a crawler refused at the edge never reads your robots.txt at all. Check by requesting your own homepage with each crawler’s user-agent and comparing the status code to what a browser gets. If a browser gets 200 and OAI-SearchBot gets 403, nothing else on this page matters yet.

Step 2: get into the sources ChatGPT already reads

ChatGPT rarely builds an answer from one company’s website. It synthesises from sources it treats as independent, which in most categories means editorial roundups, review platforms, forum threads and press.

The practical version of this is unglamorous: find the pages that already get cited for your category questions, and get into them. That is outreach, not publishing. A single inclusion in a roundup an engine already quotes will usually outperform a quarter of blog posts.

• Ask ChatGPT your category question and note which sources it cites. That is your target list, and it is finite.
• Prioritise pages that are already being quoted over pages with big traffic numbers.
• Reviews count. Recency counts more than volume.

Step 3: write passages that can be lifted

Answers are assembled from chunks, not whole pages. A paragraph that only makes sense after three paragraphs of setup is hard to quote. A paragraph that states a complete idea, names the subject explicitly rather than saying "it" or "this", and needs no surrounding context is easy.

Practically: answer the question in the first sentence under each heading, use the actual words a buyer would use, keep claims specific enough to be checkable, and avoid pronouns where the subject matters.

Step 4: make your identity unambiguous

• Consistent naming everywhere. If you are sometimes "Acme" and sometimes "Acme Inc" and sometimes "Acme.io", you are harder to resolve confidently.
• Organization and Person schema with sameAs pointing at the profiles that corroborate you.
• A clear, current statement of what you do and who you serve, in plain language, on a page that is easy to find.
• Watch out for name collisions. If your brand name is also an ordinary word, engines and your own measurement will both struggle.

Step 5: measure the same prompts every month

1. Write fifteen to twenty prompts a real buyer would type, mixing category questions, comparisons and problem-led phrasing.
2. Run them in ChatGPT on a fixed cadence and record who is named, in what order, and which sources are cited.
3. Record what is said about you, not only whether you appear. Wrong is worse than absent.
4. Watch the trend rather than any single reading. Answers vary between runs.

How the other engines differ

The mechanism is shared across engines, but each has a tilt worth knowing if you are trying to work out why you appear in one and not another.

• ChatGPT Search runs on the Bing index, so Bing indexation and Bing Webmaster Tools matter more here than most teams expect.
• Perplexity is citation-first and rewards fresh, concrete, well-sourced pages. It is often the fastest engine to reflect a new off-site mention.
• Google AI Overviews draws heavily on Google organic results, so classic SEO strength is the biggest single lever for that surface.
• Gemini shares signals with Google Search and leans on entity authority, named authors and Person schema.
• Claude favours credible, well-structured sources and clear author identity.

Optimise for the shared mechanism, not one engine

Those differences are real but second-order. Crawlable, quotable, credible and current is what all five reward, and a page that satisfies those tends to appear across the set rather than in one place.

Chasing a single engine's quirk generally wins you one screenshot rather than a position. Engines also change their weighting without announcing it, which makes anything tuned tightly to one surface a depreciating asset.

What this does not do

None of this guarantees a citation, and anyone offering that guarantee is selling something they do not control. What it does is make you the obvious thing to name: reachable, corroborated, quotable and unambiguous.

This is the method Ante runs for clients at a published $6,000 a month. It is written out in full here because a method you cannot describe is not one worth buying. If you would rather see where you stand first, the free AEO report measures these prompts across engines and returns the picture.

Frequently asked questions

How do I get my brand mentioned in ChatGPT?

Make sure ChatGPT’s crawlers can reach your site, get mentioned in the third-party sources it already cites for your category, structure your pages so a self-contained passage can be quoted, and make your identity unambiguous with consistent naming and schema. Then measure the same prompts monthly to see whether any of it moved.

Does ChatGPT read my website directly?

Sometimes, through OAI-SearchBot for its index and ChatGPT-User for live fetches when someone asks about you. But most answers are assembled from third-party sources rather than from any single company’s website, because engines weight corroboration they did not get from you. Your site is necessary and rarely sufficient.

Why does ChatGPT mention my competitors but not me?

Usually because the sources it trusts mention them and not you. Check three things in order: whether ChatGPT’s crawlers can actually reach your site, whether you appear in the roundups and review platforms it cites for your category, and whether your pages contain passages that can be quoted on their own. In most cases the gap is the middle one.

How long does it take to get cited by ChatGPT?

Third-party placements can move an answer within weeks, because engines re-read those sources often. Changes on your own site take longer to propagate. Plan on a quarter before the picture is clear, and expect the first movement to come from somewhere other than your own blog.

Does llms.txt help with ChatGPT citations?

Barely. Google has said no new machine-readable files are needed to appear in AI features, and independent monitoring found AI bots hit llms.txt on roughly 0.1% of requests. It costs an hour, so add it, but treat anyone selling it as the strategy with suspicion.

Can I pay to appear in ChatGPT answers?

Not for the organic answer, which is the point of doing this work. Advertising products exist and are expanding, but the cited brands in a normal answer are there because sources support them, not because anyone bought placement.

How do I know if my site is blocking ChatGPT?

Request your own homepage with the crawler user-agents, OAI-SearchBot and ChatGPT-User, and compare the status codes to a normal browser request. If a browser gets 200 and the crawlers get 403, you are blocked, most likely at the CDN. This is worth checking even if you never configured such a block, because several providers enable AI crawler blocking by default.