Does ChatGPT recommend lawyers? What it actually says when someone asks for an attorney

How often does ChatGPT name a specific law firm?

Most of the time, when the prompt is specific. Citorian ran 359 lawyer prompts across ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews in June 2026, covering New York, Los Angeles, Chicago, Houston and Phoenix. The engines named at least one specific firm in 62% of answers. The share moved sharply with prompt type, which is the finding that matters most for a firm deciding what to publish.

Direct vetting prompts (“is X a good firm”, “who are the best truck accident lawyers in Houston”) produced a named firm 94% of the time. Prompts naming a specific injury produced one 79% of the time. Question-led prompts (“how do I find a lawyer for…”) dropped to 23%. Situational prompts, the way a frightened person actually types after an accident, produced a named firm only 18% of the time; the rest of the time the engine gave advice and told the person to consult a lawyer, without saying which one.

How many people actually ask ChatGPT for a lawyer?

A large and fast-growing minority. iLawyerMarketing’s August 2026 survey of 1,110 US adults found that 41.9% would use ChatGPT to research which lawyer to hire, up from 28.1% in 2025, 21% in 2024 and 9% in 2023. Google was still the most-used source at 71.9%, but that figure was 86.7% a year earlier. Just over half of respondents (50.1%) said they would use at least one AI answer engine, and 9.5% said they would use only AI.

The age split is the part law firms tend to get backwards. Respondents aged 45 to 60 were the most likely group to use AI for firm research, at 76%. This is not a channel for young clients with small matters. It is the demographic that hires estate planners, litigators and injury firms.

Clio’s 2025 Legal Trends Report, published October 2025, adds the intent layer: more than half of consumers said they had used or would consider using AI to answer a legal question, and among those who had used it, 28% were directed to contact a lawyer. The AI is doing intake triage before the firm ever hears about the matter.

Which sources does ChatGPT read before it names a firm?

Directories and reviews, mostly. In the Citorian sample, Super Lawyers accounted for 33% of the sources the engines cited, Justia 20%, Google (Business Profile and reviews) 12%, Reddit 8%, Attorney at Law Magazine 7%, Best Law Firms 6%, Avvo 5% and YouTube 5%. Martindale-Avvo’s April 2026 analysis of its own query data found the same cluster for ChatGPT specifically: Super Lawyers, Avvo, Martindale-Hubbell and FindLaw were the most-cited sources.

Two things follow. First, a firm’s own website is rarely the citation; it is the corroboration. The engine names the firm because three or four independent sources agree the firm exists, practices in that area, and is well reviewed. Second, the list is not fixed. Reddit is read far more than it is cited: Ahrefs data reported by Lexgro shows ChatGPT naming Reddit in 1.93% of answers while pulling Reddit pages in 67.8% of its retrievals. What the engine reads and what it shows are different lists.

• •Super Lawyers, Justia, Avvo, FindLaw, Martindale-Hubbell and Best Law Firms: the directory layer. Completeness and consistency across these is the floor.
• •Google Business Profile: the review signal. SOCi found ChatGPT-recommended locations average 4.3 stars and that reviews act as a filter, not a ranking factor.
• •Reddit and YouTube: read heavily, cited less, and both outranked Avvo and Yelp as citations in the personal injury sample.
• •Legal press and bar records: the corroboration that separates a firm the engine trusts from one it merely knows about.

Why does ChatGPT name one firm and not a bigger competitor?

Because ChatGPT is not reading Google’s rankings. Martindale-Avvo’s analysis found ChatGPT’s answers match Google’s top ten results for legal queries less than 25% of the time, against roughly 75% for Perplexity and Claude and about 50% for Gemini. ChatGPT retrieves through Bing’s index and its own crawl, weights consistency across sources heavily, and is indifferent to ad spend. A firm on page two of Google with a complete directory footprint, steady reviews and a clear practice focus can be named ahead of a firm that outranks it.

Forward Push ran the informal version of this test, asking ChatGPT to recommend an attorney in one city and then looking at what the named firms had in common. They were not the biggest advertisers. They had dozens of pages of depth per practice area, hundreds of Google reviews, and the same identity on LinkedIn, YouTube, the directories and local news. That is the pattern Firm AEO builds toward: the engine names the firm whose record is the easiest to corroborate.

What changed in 2026 that makes this more important?

Three product changes. On March 26, 2026, ChatGPT added opt-in device location sharing, so “lawyer near me” prompts now resolve against the user’s actual location rather than a guess from context. On April 22, 2026, Best Lawyers launched a ChatGPT app that answers “find me a [specialty] lawyer in [location]” from its peer-review rankings, which puts a directory’s data directly inside the conversation. And OpenAI announced ChatGPT reached 900 million weekly active users on February 27, 2026.

The practical effect is that the local, transactional lawyer prompt, the one that used to belong entirely to Google’s map pack, is now answered inside ChatGPT with a named shortlist. SOCi’s February 2026 index measured how exclusive that shortlist is: ChatGPT recommends 1.2% of business locations, Perplexity 7.4% and Gemini 11%, against 35.9% of locations that appear in Google’s local three-pack. Being named is the exception, and it is earned upstream of the prompt.

What should a law firm do with this?

Test it, then fix what the test shows. The first step costs nothing: open ChatGPT in a private window, set your location, and ask the five questions a client in your practice area would ask, three times each. Log who gets named. Most firms find they are absent from the direct vetting prompts where the engine names someone 94% of the time, which is the highest-value gap to close.

The fix is not a ChatGPT trick. It is the record: every directory profile complete and identical, a review cadence that keeps the trailing 90 days active, one clear practice-area page per matter type written so a single paragraph answers the question, and attorney-level entities the engine can resolve. Firm AEO runs that record-building as a monthly system, measures it as an appearance rate across an agreed set of around fifteen client questions, and guarantees the firm is named across that set within 90 days or that period’s fees are waived. The appearance-rate measurement is the point: there is no rank in ChatGPT, only how often you are the answer.