Skip to main content

How to get your brand recommended by ChatGPT and AI assistants

There is no ranking to buy inside an AI assistant and no submission form. The evidence points one way: assistants name brands that are widely and credibly mentioned across the web, and that publish pages which are easy to quote accurately. Everything below follows from those two things.

What the data says correlates with being named

Ahrefs studied 75,000 brands to see which signals track with appearing in Google's AI Overviews. Branded web mentions correlated at 0.664, more than three times the 0.218 they measured for backlinks. Branded anchors came in at 0.527, branded search volume at 0.392 and Domain Rating at 0.326.

Horizontal bar chart comparing correlation with AI Overview brand visibility across 75,000 brands: branded web mentions 0.664, branded anchors 0.527, branded search volume 0.392, Domain Rating 0.326, referring domains 0.295, backlinks 0.218.
Branded mentions correlate with AI Overview visibility roughly three times more strongly than backlinks. Source: Ahrefs, 75,000 brands.

Two caveats from the authors, which we repeat because they matter: correlation is not causation, and they describe all the factors studied as showing moderate to very weak correlations. This tells you where to look. It does not promise a result.

Why your Google position will not carry you

Across 15,000 long-tail queries, only 12% of URLs cited by ChatGPT, Gemini, Copilot and Perplexity ranked in Google's top 10 for the original prompt. Ranking first and being invisible to assistants is an entirely normal state to be in.

Bar chart showing the share of AI-cited URLs that also rank in Google's top 10, by assistant: Perplexity 28.6 percent, Gemini 8.6 percent, Copilot 8.2 percent, ChatGPT in-text 8 percent, ChatGPT references 6.1 percent.
The spread between assistants is wide enough that one blended figure would hide it. Source: Ahrefs, 15,000 long-tail queries.

An ethical checklist

  • Answer the real question in the first 60 words of the page, before any preamble.
  • Use headings that match how a buyer phrases the question, so readers and assistants can find the answer.
  • Put a checkable fact in each section: a price, a date, a quantity, a named source.
  • Keep pages current and stamp them with a visible updated date.
  • Earn genuine mentions: original data, useful tools, expert commentary people want to quote.
  • Make sure assistants can read you at all, with crawlable HTML and valid structured data.

What not to do

Do not buy mentions, spin up fake reviews, or plant your brand in comment threads. Beyond being against every major platform's rules, it is fragile: these are correlational signals inside systems that are actively defended, and a brand caught manipulating them loses the credibility that made it citable in the first place.

How to know whether it worked

Ask the assistants the same unbranded questions on a schedule and record which brands they name, with the engine and the sample size against every number. Do not average across engines, and do not mix branded prompts into the discovery score. If the measurement cannot tell you it did not work, it cannot tell you it did.

Sources