How to get your brand recommended by ChatGPT: what's known and what isn't

“How do we get ChatGPT to recommend us?” is the question most marketing teams are now asking. There is a lot of confident advice about it. This guide separates what OpenAI has actually documented from what is guesswork, and lists the steps that rest on evidence.

What is known

A few things are documented by OpenAI itself.

• Its search crawler decides eligibility. OpenAI’s crawler documentation says sites that block OAI-SearchBot “will not be shown in ChatGPT search answers”. It also says a robots.txt change can take about 24 hours to take effect.
• Training and search are separate. Blocking GPTBot, the training crawler, tells OpenAI not to use your content for training. OpenAI gives the example of a site that allows OAI-SearchBot to appear in search while disallowing GPTBot.
• Search goes through outside providers. ChatGPT may share search queries with “third-party search providers such as Bing”.
• Searched answers cite their sources, with inline links and a list of sources at the end.

Google has said something similar about its own AI features: “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary,” and the usual SEO fundamentals still apply.

What isn’t known

• How ChatGPT ranks the pages it finds, or how it chooses which brands from those pages to name.
• How much the ChatGPT app differs from OpenAI’s API. The app adds memory and its own settings, so one person’s ChatGPT can answer differently from another’s.
• How often a buying question triggers a search. Across all kinds of questions, an Ahrefs study of 66.6 million responses from free, logged-out ChatGPT (August–September 2026) found about one in seven contained fan-out queries, the shorter searches ChatGPT writes for itself. For questions about what to buy, nobody has published a reliable figure for how often ChatGPT searches.

Treat any “secret ranking factor” with suspicion. If OpenAI hasn’t published it, the person selling it is inferring it.

What our runs found

Generic advice often says “get on Reddit” or “make YouTube videos”. The evidence for that comes largely from US data: Ahrefs’ tracking of US queries puts Reddit and Wikipedia at the top of ChatGPT’s most-cited domains.

Our own measurements in India point somewhere else. Citationly runs for an insurer and two packaged-food brands recorded 222 answers from OpenAI’s model through its API, with web search on, between late September and early October 2026. The most-cited websites were brands’ own sites, plus the regulator’s site for insurance. Reddit was cited in 2 of the 222 answers, YouTube in 1.

We can’t tell whether that is because of the API, the Indian market or the style of buying question. Ahrefs’ newer study does report a shift towards official sources in its US data during 2026. The lesson isn’t “ignore Reddit”. It is: look at what your own category’s answers cite before you spend money on any channel.

Steps that rest on evidence

Step	Why it matters	Evidence
Allow OAI-SearchBot in robots.txt	Blocked sites aren’t shown in ChatGPT search answers	OpenAI
Make sure Bing has indexed your site	ChatGPT search uses third-party providers such as Bing	OpenAI
Put key text in the HTML, not only in JavaScript	A Vercel/MERJ study (Dec 2024) found OpenAI’s crawlers didn’t run JavaScript (Google’s and Apple’s did)	Vercel
Have a page that answers each buyer question	Answers cite pages that state the facts the buyer asked about	What the cited pages in your own runs show
Be on the other pages your category’s answers cite	Those pages shape the answers that leave you out	Your own category’s citations

1. Make your site readable. Check robots.txt, Bing indexing, and whether your text appears without JavaScript. Our 15-minute checklist and the free checker cover this.

2. Write the page the buyer’s question needs. If buyers ask “which term plan covers critical illness?”, a page that plainly answers that, with the plan name, what it covers and the price written as text, gives an answer something to cite. Use the words buyers use, not only your internal product names.

3. Keep your facts the same everywhere. Your prices, product names and claims should match across your site, marketplace listings and comparison sites. When answers cite several pages, conflicting facts give the answer reasons to hedge or leave you out. That last part is our inference, not something OpenAI has stated.

4. Earn a place on the pages your category relies on. If a review site, a comparison page or an industry list keeps appearing in the answers that leave you out, find out whether you are on it and whether what it says is current.

Advice to treat with care

• Special AI files. Google says you “don’t need to create new machine readable files, AI text files, or markup” for its AI features. An llms.txt file is cheap to add, but don’t expect it to change what ChatGPT says.
• Channel advice without data. “Post on Reddit” or “make videos” may be right for your category or a waste of budget. Check what your answers actually cite.
• Promised results. No one can promise that ChatGPT will name you. A provider can only show you where it looks and measure again.

Measure before and after

Because answers vary from one ask to the next, a single screenshot can’t show whether anything changed. Record a baseline on a fixed set of questions, act, then ask the same questions again a few weeks later. Two runs show a difference, not proof of what caused it. Our guide to running an AI visibility audit explains how to do this fairly.

How Citationly helps

A Citationly report shows, for your category, which brands the answers name and which websites they cite when they leave you out, so your next steps come from evidence rather than generic advice. The audit includes a second run on the same questions 4 weeks later. Book a demo to see a report.