To check whether ChatGPT mentions your law firm, ask it the same questions a prospective client would type, by practice area and city, both logged in and logged out, and read exactly what it names. Run each question more than once, since ChatGPT writes a fresh answer every time rather than pulling from a saved list.
How Do You Check If ChatGPT Mentions Your Law Firm?
Open ChatGPT and ask it the kinds of questions a potential client would actually type, in plain language rather than marketing language. Ask about your practice area and your city together, for example a specific type of case combined with a specific town, and read the whole answer, not just the first line. Do this once while logged into an account and once in a private or logged-out browser window, because the two can return different results.
Write down exactly what ChatGPT says: whether it names your firm, what it says about the firm, which other firms or lawyers it names instead, and whether it shows any source link (sometimes called a citation). Our FAQ on whether ChatGPT cites law firm websites covers when and why a source link does or doesn't appear. That single check, repeated with a handful of realistic questions, is the whole method. The rest of this guide makes that check more useful and repeatable.
What Prompts Should You Test First?
Start with a short, fixed list of prompts rather than typing whatever comes to mind each time, so later checks are comparable to earlier ones. Three kinds of questions are worth testing first.
Practice area plus location questions, phrased the way a client would say them, such as a specific type of case combined with your city or region. "Best" or "top" questions, such as asking ChatGPT to name a good lawyer for a specific situation, since these are the questions most likely to produce a named recommendation rather than general advice. Direct comparison questions that name a competitor firm you already know of and ask how it compares to firms in the same area, since these reveal who ChatGPT considers when a name is already on the table.
Keep the wording of each prompt exactly the same every time you retest it. Changing even a few words can change the answer, which makes it hard to tell whether a real shift happened or you simply asked a different question.
Why Does ChatGPT Give a Different Answer Each Time?
ChatGPT does not keep a fixed list of law firms that it checks against. It generates a new answer, piece by piece, each time you ask, based on patterns learned during training and, in some cases, information it looks up while answering. That means the same question asked twice can produce two different answers, even within the same day.
One good answer that names your firm is not proof that ChatGPT reliably recommends you. One bad answer is not proof that it never will. Run the same prompt several times, on different days, before drawing any conclusion about a pattern. A single result is a data point, not a verdict.
How to Keep a Record of What ChatGPT Says About Your Firm
A one-time check tells you what ChatGPT said on one day. A log tells you whether that is changing. Keep a simple spreadsheet, one row per check, with the same fields every time.
Add a new row every time you or someone on your team runs a check, using the same prompts from your list. After a run of entries over time, you will be able to see whether your firm is showing up more consistently, less consistently, or about the same, instead of relying on memory of a few scattered checks.
Should You Track Gemini and Perplexity Too?
Yes. Clients who ask ChatGPT a question about finding a lawyer are also asking Google Gemini and Perplexity the same kind of question, and each of these tools can give a different answer, because each is built differently and draws on different information. Checking only ChatGPT gives you one piece of a bigger picture.
The underlying work of making a firm easier for any of these tools to describe accurately is often called generative engine optimization (GEO) or answer engine optimization (AEO), depending on which engine is under discussion. Both terms cover the same general goal: making your firm's information clear enough for an AI tool to use correctly when it forms an answer. Our guide to generative engine optimization covers that work in more depth, and the FAQ on whether ChatGPT can recommend a lawyer covers what ChatGPT specifically does and does not do with legal recommendations.
What Makes ChatGPT More Likely to Mention Your Firm?
An AI tool can only describe your firm using information it can find and trust. If your website states plainly, in ordinary sentences, what areas you practice, where you are located, and who you represent, that gives any answer engine, ChatGPT included, something concrete to work with. Vague pages, or pages where that information is buried in images or navigation menus rather than in text, give it less to draw on.
Structured data is a standardized way of labeling facts on a webpage, using a specific code format, so that computer systems, including search engines and AI tools, can read those facts directly instead of inferring them from paragraphs. Google's own search documentation describes how this works for a business with a physical location, and Schema.org, the shared vocabulary that search engines and AI tools draw on for this kind of labeling, defines a LegalService type built for firms like yours. It lets a page state its name, address, and practice areas in a form software can parse without guessing.
A newer, related idea is llms.txt, a plain text file some sites are starting to publish to give AI systems a clear, direct summary of what the site is and contains. It is a proposed, community-maintained format. No major AI provider has confirmed it reads this file, so treat it as a reasonable thing to have in place rather than a guaranteed lever.
None of this guarantees a mention. It gives ChatGPT clearer material to work with, which is the most that any firm actually controls. If your firm never appears no matter how many prompts you try, that is a different, more specific problem; see why your law firm might not be appearing in AI search results.
When Is Manual Checking Not Enough?
Manual checks work well for a first look, and cost nothing beyond your time. They also have limits. Running the same set of prompts across ChatGPT, Gemini, and Perplexity on a regular schedule, by hand, is easy to plan and easy to let slide after the first few weeks. It is also hard to keep consistent if more than one person on your team is doing it.
If you want a current, fuller picture without maintaining that manual log yourself, a free AI visibility check shows what AI tools currently say about your firm across engines in one pass. It is a reasonable next step once manual spot-checks have told you enough to know the question matters for your firm.