How ChatGPT Chooses Which Attorneys to Recommend
A prospective client asks ChatGPT "who is the best divorce attorney near me for a high-conflict case," and within seconds gets two or three names, often with a short rationale attached. That client rarely explores five more law firm websites afterward. They contact one of the names the AI already screened for them. This guide explains exactly how AI engines make that selection, and what a firm needs to do to be one of the names.
AI Does Not Rank Firms the Way Google Does
Google ranks pages by relevance and authority signals accumulated over years, which is exactly why legal directories with decades of backlinks dominate page one for almost every practice-area query. AI engines work differently. When asked a recommendation question, they retrieve current content from across the web, evaluate which sources are trustworthy and machine-readable, and generate an answer that names specific firms, not a list of links to click through.
The Four Conditions a Firm Must Meet
- The AI can read the site. If
GPTBot,PerplexityBot, orGoogle-Extendedis blocked inrobots.txt, or if practice-area details live only behind a JavaScript-rendered widget, the AI cannot retrieve the content at all, regardless of how good it is. - The AI can classify the firm. Without
LegalServiceorAttorneyschema, the engine has to infer practice areas and location from marketing prose, which it does unreliably. Explicit markup removes the guesswork. - The AI can extract facts. A biography page that reads "passionate about fighting for justice" gives the AI nothing to cite. A page stating "handled over 200 personal injury cases in Brisbane since 2014, with a combined settlement value exceeding $40 million" gives it a citable fact.
- The AI can verify trust. Bar admission, years in practice, and case outcomes need to be consistent across the firm's own site, its directory listings, and review platforms. Contradictions between sources reduce confidence in all of them.
Why the Same Two or Three Firms Keep Getting Named
AI citation compounds. Once an engine has cited a firm reliably for a given query, subsequent retrievals tend to reinforce that pattern unless a better-structured competitor displaces it. This is why firms that move early on GEO tend to keep their position, while firms that wait find it progressively harder to break in for the queries that matter most to their practice.
Six Practical Fixes That Move the Needle
| Fix | What it changes |
|---|---|
Deploy LegalService and Attorney schema | Gives the AI an unambiguous, structured description of the firm and its attorneys |
| Rewrite attorney bios as fact-first | Replaces vague marketing language with specific, extractable achievements |
| Build dedicated practice-area and location pages | Lets the AI match the firm to specific, high-intent queries instead of generic ones |
| State bar admission and credentials in text | Makes trust signals machine-readable, not just visually present |
| Align directory listings with the firm's own site | Removes contradictions that undermine AI confidence |
| Publish substantive case-result summaries | Gives the AI genuinely citable content beyond a services list |
See exactly what ChatGPT says about your firm today
We test whether ChatGPT, Perplexity, and Google AI recommend your firm for the queries your prospective clients actually use, free, in a 30-minute plain-language call.
Get My Free AI Visibility Audit →For the complete strategic framework, read the complete 2026 guide to AI search optimization for law firms. Related guides: the death of traditional legal SEO, personal injury firms, the statistics.
Sources: Haute Lawyer Network and 5W, "The 2026 Legal AI Visibility Report."