How to rank in ChatGPT
ChatGPT does not rank you the way Google does. It answers the question the buyer asked with a few names, and nothing else. Your firm is either named or it is not. Here is how it picks, what it relies on for legal questions, and what a law firm can do to be one of the sources it trusts.
A general counsel has your name from a trusted referral. Before she calls, she opens ChatGPT and asks whether your firm has handled her kind of matter, who the partner would be, and whether anything should give her pause. Your firm either holds up through all four answers or it does not. Nobody at your firm will ever see that conversation.
Your next client asks AI who to hire. Your current clients ask AI whether to keep you.
There is no ranking to win
You do not rank in ChatGPT. It gathers sources, weighs them, and names a few firms. The job is to be one of the sources it trusts.
Google gives the reader ten links and lets them choose. ChatGPT gives an answer and has already chosen. Either your firm is among the sources it pulled and believed, or the answer is written without you. Keywords and links are Google's levers. A model moves on whether your firm is in the sources it reaches for, described in a way it can use, and backed up by sources it already trusts.
What ChatGPT cites
The directory law firms invest in most barely registers with ChatGPT. The source a lawyer would never cite in a brief is the one it leans on.
Perplexity, another engine your clients use, builds its legal answers around a different directory entirely, Super Lawyers. There is no one AI to be visible to. Your clients use five engines, ChatGPT, Perplexity, Gemini, Claude, and Google's AI Overviews, and each builds its answer from different sources. Your firm has to be measured on each one separately.
Why the answer moves
The same question, asked three times, does not return the same firms. In a 1,600-answer Perplexity study, about half the firms named showed up in only one of three identical runs (InterCore, 2026).
The list you saw this morning is one snapshot. The list your client sees tonight may be another.
A single check tells you where you were. What matters is the trend: how often your firm appears, on which engines, and whether that is rising. That means measuring multiple times, every month against where you started, not one report. Read how that works in the Visibility Loop.
What helps
Three things decide whether ChatGPT finds and trusts your firm.
Be the source. When an engine describes your firm from a directory profile or a forum thread, it is working from someone else's words, which may be old, thin or wrong. Give it your own practice page with specifics: the dispute types, the industries, the courts, the deal sizes. "We handle complex commercial disputes" gives a model nothing to use.
Write pages it can lift an answer from. An engine lifts passages rather than reading the whole page. First sentences that answer the question, headings that match what a client would ask, facts in short paragraphs. Key facts buried in the fourth paragraph of a partner biography get skipped. How each engine reads a page is in AEO explained; the difference between writing for search and writing for answers is in the acronym guide.
Be present where that engine looks. ChatGPT leans on Wikipedia and the open web. Perplexity leans on Super Lawyers. Effort poured into a prestigious directory the engine barely consults is wasted. See which sources feed the answers about your practice on each engine, and fix the ones feeding answers about someone else.
What is different for a law firm
For a law firm this runs at three levels, the firm, the practice and the lawyer. A firm can be named while its practice is not, and a practice can be named while the partner who built it is missing from every answer.
When a client asks ChatGPT who handles cross-border restructuring, the answer is a name, or three names, with a firm attached. The work follows that name. If the partner who has done this for twenty years has no page that describes her matters in plain terms, the model names someone else, and your firm never learns why the call did not come. Does AI know your best partner exists?
So the starting point is a measurement: on each of the five engines, at the firm, practice and lawyer level, on the questions your clients ask, which sources feed the answers and who is named. If you are choosing how to get that measurement, the questions to put to any vendor are in how to choose an AI visibility platform.
We don't ask you to believe us. We show you your own result.
Frequently asked questions
Not the way a page ranks in Google. ChatGPT keeps no list of firms. For each question it gathers sources, weighs them, and names a few. A firm is named when it is in the sources the model trusts, on each engine, at the firm, practice and lawyer level.
In a Firesign audit of legal answers, ChatGPT cited Wikipedia in 35.6% of answers and Chambers in 4.1%. Perplexity cited Super Lawyers in 50.7%. Each engine draws on different sources, so each has to be measured separately.
Because it builds each answer fresh from many sources every time. In a 1,600-answer Perplexity study, about half the firms named showed up in only one of three identical runs (InterCore, 2026). A single answer is a snapshot. Track how often your firm is named, on each engine, month over month.
Be the source: publish specific pages on what you do, for whom, and what the matters look like. Write pages an engine can lift an answer from: lead with the answer, use headings that match clients' questions, keep paragraphs short. Be present where that engine looks: fix the sources feeding the answers. And measure at three levels, because the partner is often the name being asked for.
Is your firm one of the sources ChatGPT trusts?
The free, live check shows you where you stand across ChatGPT, Perplexity, Gemini, Claude, and Google's AI Overviews, which sources are feeding the answers, and which firms are on the shortlist in your place.
Run your AI visibility check