How law firms get recommended by ChatGPT

ChatGPT recommends the law firms it can find, read and confirm. When someone asks for a lawyer, it searches the web, reads your site alongside directories, reviews and press, and names the firms those sources agree on. A firm gets on that short list by letting the engine crawl its site, stating plainly who practices what and where, and making sure outside sources say the same thing.

Ask ChatGPT "who should I call about a car accident in my city" and you get a few firm names with a reason for each, not ten blue links. For a growing number of prospective clients, that answer decides who gets the intake call.

How many prospective clients ask ChatGPT for a lawyer?

Enough that it now shows up in consumer surveys as the second source after Google. In iLawyer Marketing's 2026 survey of 1,110 U.S. adults aged 18 to 65, 41.9% said they would use ChatGPT if they needed to research a lawyer for an important legal issue, up from 28.1% in 2025. Google still led at 71.9%, down from 86.7% the year before.

Two caveats before you act on that. iLawyer Marketing sells marketing to law firms, and the survey asks what people would use, not who they actually hired. The same study also found that just over 70% of ChatGPT users would still check Google, down from over 94% in the 2025 edition. Read it as direction, not as a share of your signed cases: most people still cross-check, but fewer than before.

How does ChatGPT decide which firms to name?

It runs a search, reads a set of pages and writes an answer from what those pages say. OpenAI documents that its search crawler, OAI-SearchBot, is what makes a site eligible, and that sites which opt out of it will not be shown in ChatGPT search answers (they can still appear as navigational links). OpenAI does not publish its ranking factors, so anyone who claims to know the exact formula is guessing.

What you can see is the output. When ChatGPT searches, it links the pages it drew on, and for a legal question those can include firm websites, legal directories, review profiles, local news and forum threads. The firms that get named tend to be the ones several of those sources describe the same way. That is why a firm with an expensive website can be absent while a smaller firm with accurate directory profiles and detailed reviews gets named.

Google works on the same principle. Its documentation says AI Overviews and AI Mode may use a "query fan-out" technique, issuing several related searches to build one response, and that there are no extra requirements to appear beyond being indexed and eligible for a snippet in regular Search. Good search hygiene is the entry ticket to both engines.

Which sources do AI engines read about a law firm?

Four groups of sources carry most of the weight for legal questions, and you control some of them more than others.

The pattern to notice: the less you control a source, the more it confirms what your own site says. Engines treat your website as a claim and the other three rows as evidence.

What makes an engine pass over your firm?

Usually one of three things: it can't read you, it can't tell what you do, or outside sources disagree with you.

It can't read you. A firewall or CDN bot rule can block OAI-SearchBot without anyone at the firm knowing. OpenAI recommends allowing the crawler in robots.txt and allowing requests from its published IP ranges. Pages that only render through heavy scripts, or practice areas buried in PDFs, are harder for any crawler to use.

It can't tell what you do. A homepage that says "aggressive advocates for justice" tells an engine nothing it can quote. "We represent injured people in car, truck and motorcycle cases in two counties, on contingency" gives it a fact to repeat. Attorney bios that skip bar admissions and practice areas leave the same gap. Structured data helps here: Schema.org's LegalService type describes a law firm as a local business, so your name, address, practice areas and attorneys become machine-readable rather than inferred.

Outside sources disagree. A partner who left two years ago still listed on a directory, an old suite number, a practice area you dropped: each one gives the engine a reason to hedge, and a hedging engine names someone else. Google's own Business Profile guidelines add a trap specific to firms: lawyers are individual practitioners, and when several attorneys work at one office, the firm should have its own profile and each attorney's profile should carry only their name. Keyword-stuffed names like "Smith Law Car Accident Lawyer" are not permitted and can get a profile suspended.

Do the advertising rules apply to what AI repeats about you?

Yes, because the engine is quoting your communications. Your website, bios, profile descriptions and the review requests you send are all communications about your services under your state's version of ABA Model Rule 7.1. Three points matter most when you write copy an engine will lift.

First, results. The ABA comment on Rule 7.1 says a truthful report of past results can still be misleading if it creates an unjustified expectation that other clients will get the same result. An engine summarizing your verdicts page strips away your disclaimer, so put the qualifying context in the same sentence as the result.

Second, specialist claims. Model Rule 7.2(c) bars stating or implying certification as a specialist unless an approved or ABA-accredited body certified you and the communication names that body. "Specialist" in a bio is a phrase an engine will happily repeat.

Third, reviews. Rule 7.2(b) bars giving anything of value for recommending your services, beyond narrow exceptions such as nominal thank-you gifts. Ask satisfied clients for honest reviews; don't pay for them or trade discounts for them. Your state's rules may be stricter than the model rules, so check your bar's advertising rules before you change review practices.

Steps you can take this month

None of these need a vendor. A managing partner and an office manager can do most of them in a few afternoons.

1. Check the crawler. Open yoursite.com/robots.txt and confirm OAI-SearchBot is not disallowed. Ask whoever manages your hosting or CDN whether bot protection blocks it.
2. Rewrite your first paragraph. On the homepage and each practice page, state who you represent, which case types, where, and how you charge, in plain sentences.
3. Fix attorney bios. Each bio should list bar admissions, years in practice, practice areas and office. Remove anyone who has left.
4. Add LegalService structured data with your firm name, address, phone and practice areas, and make it match your Google Business Profile exactly.
5. Audit five directory profiles. Claim and correct Avvo, Justia, FindLaw, Martindale-Hubbell and your state bar listing so the name, address, phone, attorneys and practice areas match your site.
6. Ask for specific reviews. Ask recent clients to describe the type of matter and how the firm communicated, within your state's testimonial rules and with nothing offered in return.
7. Answer the questions your intake team hears. Write one page per question ("how much does a divorce lawyer cost here," "what happens after a DUI arrest") that opens with the answer. Our breakdown of what law firm SEO costs in 2026 covers what to expect if you hire this out.

How to check where you stand

Pick the five questions your best clients ask before they call, phrased the way they would type them, with your city. Ask each one in ChatGPT, Perplexity and Google's AI Mode. Write down which firms are named and which sources are cited. If you are missing, the cited sources are your to-do list: those are the directories, articles and threads the engine trusts for your practice area.

If you'd rather have it done properly, our free AI visibility audit runs your real practice-area questions across five engines and shows which firms are named instead of you, and why. More on how we work with firms is on the law firms page.

Frequently asked questions