How are potential clients using AI to find lawyers?

Q: How are potential clients using AI to find lawyers?

Mostly as a research and triage tool before they call anyone. Clients ask ChatGPT, Perplexity, Gemini, Claude, and DeepSeek about their situation, what kind of lawyer they need, what questions to ask, and only then for firm names. The engines disclaim, push toward Avvo and state bar referrals, and name firms cautiously, usually only on a follow-up query.

Q: Will a client hire a lawyer directly from an AI answer?

Rarely in 2026. AI engines are a triage and credibility-check layer. Most hires still come from Google Maps, referrals, or existing relationships. AI visibility is a confirmation signal at the selection step, not a primary acquisition channel yet.

Q: Which engine is most willing to name specific firms?

Perplexity is most willing to name firms on the first response. ChatGPT waits for an explicit follow-up. Gemini mirrors Google's local pack. Claude is the most conservative and often declines to name firms at all, redirecting to bar referral services. DeepSeek is uneven across runs.

Q: What signals make a firm appear in the shortlist when an AI engine is asked for specific firms?

A complete Avvo profile with peer endorsements, Super Lawyers selection, published case results within ABA Model Rule 7.1 limits, substantive Google reviews mentioning the practice area and city, and long-form practice-area pages on the firm site.

Question

How are potential clients using AI to find lawyers?

Maciej Grabek · Accepted Answer

Potential clients are mostly using AI engines as a research layer before they call anyone. They ask ChatGPT, Perplexity, Gemini, Claude, or DeepSeek what their situation looks like, what kind of lawyer they need, what questions to ask, and only then for firm names. The engines answer with disclaimers, push toward Avvo and state bar referrals, and name firms cautiously. The implication for firms: visibility is won upstream of the firm-name query. The shape of a real client journey in 2026 In our work auditing law firm websites and tracking AI engine output, the pattern that comes up over and over looks like this. A potential client lands on an AI engine with a problem, not a firm name in mind. They might type: "I was rear-ended at a stoplight in Houston, no injuries yet but my neck is sore. Do I need a lawyer?" "My ex stopped paying child support, what are my options in Florida?" "Got a DUI in Phoenix last weekend, first offense, what should I do?" "How do I contest a will in California if my brother was the executor?" None of these are firm-name queries. The engines treat them as situational triage. They explain the legal landscape, name the type of lawyer to look for (personal injury, family law, criminal defense, probate litigation), point to free or low-cost resources first, and only then surface specific firm names if pushed. What the engines actually say at each step Run the rear-ended-in-Houston prompt across ChatGPT, Perplexity, and Gemini. The output is not identical, but the structure converges: Disclaimer about not being a lawyer. Practical first steps: see a doctor, document the scene, file the claim, do not sign anything from the other insurer. When to consider hiring a lawyer ("if injuries develop or the insurer refuses fair settlement"). How to find one: state bar referral, Avvo, Justia, Super Lawyers. Sometimes named firms, especially if the user pushes with "can you suggest specific firms". Perplexity is the most willing to name firms in the first response. ChatGPT tends to wait until the user explicitly asks. Gemini leans hardest on Google's local pack, so its firm suggestions match what would appear in a Maps search. Claude is the most conservative, often declining to name firms at all and recommending the bar referral service. DeepSeek is uneven and inconsistent across runs. The follow-up query is where firms get picked The interesting query is the second one. After the engine has explained the situation, the client asks: "Can you suggest some specific firms in Houston with strong personal injury practices?" That is the prompt that decides whether your firm is on the shortlist or not. The signals the engines lean on for that answer: Avvo profile with a high rating, peer endorsements, and review depth. Super Lawyers selection, especially "Rising Stars" and the main list. Case results published on the firm site with concrete language about outcomes. Substantive Google reviews mentioning the practice area and the city. Long-form practice-area content on the firm site that the engine has indexed. Firms that miss the shortlist tend to share the same gaps: thin Avvo profile, no Super Lawyers presence, generic practice-area pages, and reviews that are mostly star ratings without substantive text. What clients are not asking AI for Worth saying clearly: AI engines are not yet a major source of cold legal hires. Most clients still use them as a triage and research tool, then fall back to Google Maps, a referral from a friend, or an existing relationship for the actual hire. The engines themselves push users toward bar referral services and Avvo for the actual selection step. So a firm betting that AI search will replace its referral and Google Ads pipeline in 2026 is betting too early. The right framing: AI visibility is a credibility signal at the moment of selection, not a top-of-funnel acquisition channel - yet. Three things to do now Run the four sample prompts above (or your own equivalents in your practice area and city) across all five engines. Note which firms get named in the second-query "can you suggest specific firms" follow-up. That is your real competitive set in AI search. Audit your firm's Avvo, Justia, FindLaw, Super Lawyers, and Best Lawyers profiles for completeness. Reviews with substantive text matter more than star count alone. Test what your own firm is cited for. If you are mentioned, what context does the engine put around the mention? If you are not mentioned, what is the gap? You can scan your own firm in two minutes to see exactly which of these queries surface you and which do not.