GEO for law firms: get cited when clients ask AI for a lawyer

How clients look for lawyers now

The search for a lawyer has moved into the chat box. iLawyerMarketing's 2026 consumer study puts ChatGPT at 41.9% as a research source for finding legal help — up from 28.1% a year earlier, second only to Google at 71.9%, and ahead of every directory, referral network and social platform. The pattern matters as much as the number: prospective clients rarely start by asking for a firm by name. They ask about the problem — whether an injury claim is worth pursuing, what a divorce costs, whether a first DUI needs counsel — and the engine's answer frames who they contact next.

When the question becomes 'who should I hire', the engine returns a short list of firms with reasons attached. There is no page two of an AI answer. A firm that isn't named hasn't been ranked lower; it has been left out of the consideration set entirely, and no intake dashboard records the case that never called.

The economics make this sharper for legal than for almost any other vertical. On contingency work, a single signed case can be worth the entire year's marketing budget; on hourly work, matters compound into multi-year relationships. When the margin between named and absent is one strong personal-injury case a quarter, visibility in the answers clients actually read stops being a nice-to-have and becomes intake infrastructure.

What gets a law firm cited

Legal questions sit squarely in the category engines treat most cautiously, so the sources behind lawyer recommendations skew toward bar directories, established review platforms and pages that demonstrate verifiable expertise. In practice, four things do most of the work:

• Answer-first practice pages. Each practice-area page opens with a direct 40–60 word answer to a question clients actually ask — what a case is worth, what representation costs, how long it takes — written so an engine can lift the passage whole. This is extraction, and most firm sites are structured against it.
• Attorney entities, not bio pages. Attorney and LegalService structured data with bar admissions, practice areas and office locations makes each lawyer a recognizable entity engines can attach facts to.
• A coherent directory footprint. Avvo, Justia, FindLaw and your state bar directory recur in the citations behind 'best lawyer' answers. Listings with mismatched names, addresses or practice areas dilute the signal engines are trying to confirm.
• Concrete answers to fee and process questions. Most firms publish vague reassurance about fees. Engines cite the pages that state ranges, factors and process plainly — which is also what clients want before they call.

What we'd audit first for a law firm

The first pass on any firm follows the same sequence, because failures compound in this order:

1. Crawler access. Whether OAI-SearchBot, PerplexityBot and the other AI search crawlers can reach the site at all — law-firm sites behind aggressive bot blocking are more common than you'd expect.
2. A baseline prompt set. Twenty-five to forty prompts covering your practice areas and cities, run across six engines before anything changes, so every later claim of progress is measured against a number.
3. Practice-area page structure. Whether each page contains a passage an engine could quote as the answer, or opens with 'At Smith & Associates, we understand…'.
4. Entity and directory consistency. The same firm name, addresses, attorneys and practice areas everywhere engines look, including the state bar record.
5. Competitor citation trace. Which sources engines actually cite when they name rival firms in your market — that list is the off-site roadmap.