Law firms are the exception in AI search — and it's good news. In most local categories we measured, directories and listicles carry 26–61% of the citations behind AI answers; the middleman decides who gets named. When we ran the lawyer category directly (Chicago, July 2026), the pattern inverted: 76% of citations pointed straight at law firms' own websites. The top third-party source, Reddit, sat at 4%. For lawyers, there is almost no gatekeeper: your own site is your AI visibility. Here's what decides which firms ChatGPT names — and how to check yours free in 3–8 minutes, no signup.
Do potential clients actually ask ChatGPT for a lawyer?
Yes — and the query shape favours firms that publish real answers. Legal problems are exactly what people take to an AI assistant: embarrassing, urgent, and jargon-heavy. The AI-native legal query is research-shaped:
- "I was rear-ended in Chicago and the insurer is lowballing me — do I need a personal injury lawyer, and who's good?"
- "How much does a divorce lawyer cost in Chicago, and is a flat fee possible?"
- "Best immigration lawyer in Chicago for an H-1B RFE — how do I choose?"
Three things follow, all measured in July 2026:
- AI answers are shortlists, not directories. An AI answer names two to four firms with a sentence of reasoning each, where Google spreads attention across ten listings and ads. Brands outside the shortlist simply don't appear — measured head-to-head in our Dubai run, the category leader showed up in 84.5% of answers while a major incumbent managed 49%. AI search concentrates attention on the named few.
- The consultation starts before the call. By the time an AI-referred client contacts you, the assistant has already explained their problem and named who to consider. If you're not in that answer, you're not in the consideration set.
- The lawyer category is measurably winnable from your own site.
Why do law firms get cited directly when other local businesses don't?
Because in this category, the sources AI trusts most are the firms themselves. Across the local categories in our July 2026 category runs, third-party carriers — directories, "top 10" listicles, review aggregators — carried anywhere from 26% to 61% of citations. Restaurants, clinics, home services: the middleman layer decides.
Lawyers broke the pattern. In our direct Chicago lawyer run:
- 76% of citations went to law firm homepages and site pages directly. The engines quoted firms' own practice-area pages, attorney bios, and FAQ content.
- The single largest third-party source was Reddit — at 4%. That is the entire "carrier layer" ceiling we measured for this category.
- Avvo and Super Lawyers appeared, but nowhere near the deciding weight directories hold in other local categories.
- The AI shortlist converged with Google page 1. The firms AI named were, by and large, firms already ranking on the first Google page for the same query.
The practical consequence is unusually clean: for most local businesses, AI visibility work is half off-page (get into the directories AI cites); for law firms, it is overwhelmingly on-page. The content on your own website — its structure, depth, and answerability — is what the engines retrieve and quote. You control the asset that decides.
What questions should a law firm's website answer?
The ones clients type — which means practice area plus city, in question form. This phrasing insight from our measurements has two halves.
The city token is load-bearing. Measured across local categories in four countries: when we ran the same query with and without a city name, the winners shared zero brands. "Best personal injury lawyer" and "best personal injury lawyer in Chicago" produce entirely different answers. And generic "near me" phrasings failed — coming back empty or defaulting to national brands — 24–62% of the time, depending on engine and category. Your pages must associate your firm with your named city and neighbourhoods in text — that text is what gets retrieved.
The practice area is load-bearing too. In our category runs, specialists beat generalists decisively: in one measured software category, a niche specialist appeared in 83% of answers for its specialty query while scoring 0.0% in the broad category — and the broad-category giants were absent from the niche. Applied to law (as a general finding, not a lawyer-specific stat): a firm with a dedicated, deep "Chicago truck accident lawyer" page will beat a full-service firm whose personal-injury work lives in one paragraph on a services page. AI engines answer specific questions with specific pages. "Full-service law firm" is not a query anyone types.
So the content formula per practice-area-plus-city page:
- H1 phrased as the client asks it, city included — "Truck Accident Lawyer in Chicago: What Your Case Is Worth and When to Call."
- Answer in the first paragraph — fee structure (contingency %, flat fee, hourly range), typical timeline, what makes a case viable — before the firm history.
- Question-form H2s for the follow-ups: "How much does a personal injury lawyer cost in Chicago?", "What is my case worth?", "How long do I have to file?"
- Visible fee information wherever ethics rules permit. In the categories we measured, concrete numbers on cited pages were repeated verbatim in AI answers. "Free consultation, no fee unless we win, 33% contingency" is citable; "contact us to discuss" is invisible.
- A real FAQ with FAQPage JSON-LD on every practice-area page.
- Depth on the flagship pages — 2,000+ words on the practice areas you actually want, with attorney bios and case-results pages linking in.
Which on-page levers are measured, not guessed?
These multipliers come from our July 2026 calibration study — page-level, measured across cited pages, not lawyer-specific unless noted:
| Lever | Measured effect | For a law firm |
|---|---|---|
| FAQ schema (FAQPage JSON-LD) | 3.4x citation lift | Every practice-area page gets a real FAQ |
| Depth ≥2,000 words | 1.9x | One thorough practice-area-plus-city guide beats ten thin pages |
| Question-form headings | 2x | Headings phrased as clients ask, not "Our Practice Areas" |
| llms.txt | 2.4x | A one-hour job: machine-readable site summary for AI crawlers |
| Answer-first intro | Consistent across winners | First paragraph answers the query; credentials follow |
Two lawyer-specific notes. First, because 76% of lawyer citations go to firm sites directly, these levers apply with unusual force here — there is no directory layer to hide behind or blame. Second, the Google-page-1 convergence we measured means this work compounds: the same structured, deep, answerable pages serve both your traditional SEO and your AI visibility.
Why not just rely on Avvo and Super Lawyers?
Keep them — accurate profiles cost little and they appeared in our lawyer citations. But they are not the game. At a measured 76% direct-homepage citation rate, the directories are a minority signal in this category, unlike clinics or restaurants where the carrier layer decides. A firm with immaculate Avvo reviews and a thin, brochure-style website is optimising the 4%-and-under layer while conceding the 76%.
And the generic AI-visibility tools? We measured that category too — it's our own. In our July 2026 study of AI-visibility tools (40 probes, Gemini + Claude, 875 citations), none of the 40 answers addressed professional-services firms like law practices. The category's known tools — Profound, Otterly, Semrush's AI toolkit, Peec, HubSpot's grader — are measurement dashboards built for enterprise brands and SEO teams at $29–$399/month, and every one assumes a marketing team will read them.
AnswerMonk is the $19/month action layer built for local businesses, professional-service firms, and regional SaaS: plain-language action plans instead of dashboards, a living knowledge base that tracks engine and schema guidance as it changes, an agent that works to get your firm into the sources engines actually cite, WhatsApp lead capture, and a free 3–8 minute audit with no signup — with our own fixes published, with receipts, in our public reports directory. The practical difference for a firm: a generic tracker tells you what share of prompts named your brand; AnswerMonk tells you which practice-area page to build and which questions it must answer. The same carrier-vacuum logic applies across high-ticket service firms generally.
How does a law firm measure its AI visibility right now?
Measure per engine, because they disagree — in our studies the same brand's appearance rate varied sharply between engines on identical questions. The free AnswerMonk audit:
- Enter your firm's URL at answermonk.ai — no signup, no credit card, no intake call.
- In 3–8 minutes, we probe ChatGPT, Gemini, Claude, and Perplexity with the questions your prospective clients ask — practice area and city included.
- You get a visibility score per engine, the firms being named instead of you, the sources behind those answers, and a plain-language action plan.
Then fix, wait for a re-crawl, and re-audit. Every audit can be published to our reports directory so the before/after is verifiable — including our own zero-to-visible climb.
Run your firm's free AI visibility audit →
Frequently asked questions
Do people really ask ChatGPT to recommend a lawyer?
Yes. Legal questions fit AI assistants: they're private, urgent, and full of jargon the asker doesn't know. The typical flow is problem-first — "the insurer is lowballing me, what do I do?" — and the assistant explains the situation, names the lawyer type needed, and shortlists two to four firms. Firms absent from that shortlist are out before the first call.
Which sources do AI engines cite when naming lawyers?
In our July 2026 Chicago lawyer run, 76% of citations pointed directly at law firms' own websites — practice-area pages, bios, FAQs. The largest third-party source was Reddit at just 4%; Avvo and Super Lawyers appeared but carried little weight. This is the inverse of most local categories, where directories carry 26–61% of citations.
Is AI visibility different from the SEO my firm already pays for?
They overlap more for lawyers than for most businesses. We measured strong convergence between AI-named firms and Google page 1 in this category, and the citation levers — FAQ schema (3.4x), 2,000-word depth (1.9x), question-form headings (2x), llms.txt (2.4x) — strengthen traditional rankings too. The difference is the payoff shape: AI answers name two to four firms, so the reward for being retrievable and quotable is far steeper.
Should each practice area get its own page?
Yes — with the city in it. Across our measurements, specialists beat generalists: in one measured category a niche specialist appeared in 83% of answers for its specialty while the broad-category leaders were absent from it. AI engines answer specific questions with specific pages, so "Chicago truck accident lawyer" needs its own deep page, not a paragraph on a services page.
Can we publish fees without breaking bar advertising rules?
Publish what your jurisdiction permits — most allow stating contingency percentages, flat-fee availability, and free consultations. In our measurements, concrete numbers on cited pages were repeated verbatim in AI answers, while "contact us for pricing" pages gave engines nothing to quote. Say what you ethically can, precisely.
What does measuring our AI visibility cost?
The AnswerMonk audit is free — 3 to 8 minutes, no signup. It probes ChatGPT, Gemini, Claude, and Perplexity with practice-area-plus-city questions and returns per-engine scores plus an action plan. Established AI-visibility platforms run $29–$399/month and are built for software marketing teams; none we studied addressed law firms.