GEO: How Law Firms Get Recommended by ChatGPT and Perplexity

The New Referral Nobody Is Tracking

Ask ChatGPT for the best personal injury lawyer in your city, and it will often name real firms. Ask Perplexity the same thing and you will often see the same behavior in a different interface: a short list, presented with confidence, that can read like a referral from a knowledgeable friend.

Firms on those lists are in the conversation. Firms off them may never know it happened. The behavior is measurable: a June 2026 five-engine study by Citorian, a law firm SEO and GEO agency, logged 359 AI answers about personal injury lawyers in five large US metros and found the engines named at least one specific firm 62% of the time overall. Phrasing mattered: the rate rose to 94% when the question was phrased as "best personal injury lawyer in [city]" but fell to 18% when it was phrased the way an accident victim might ask ("I was in a car accident, who do I call"). The study covered one practice area in large markets and comes from a vendor that sells this work, so other practice areas and smaller markets may differ. Generative Engine Optimization (GEO) is the practice of earning those recommendations deliberately.

How AI Engines Decide Which Firms to Name

The engines don't publish how they choose, but AI recommendations draw on information from across the web. The factors practitioners most often point to:

• Entity clarity: making your firm easy to identify as a distinct entity (name, location, practice areas), consistent across your site, Google Business Profile, bar listings, and legal directories.
• Corroboration: consistent presence across independent sources (directories, reviews, local press, "best lawyer in X" lists) appears to matter. In the Citorian study, Super Lawyers and Justia were the sources cited most often by the engines that exposed their source links, and the most-named firms had heavy presence in those directories.
• Citable content: the same answer-shaped content that powers AEO also gives engines a record of what your firm handles.
• Machine-readable structure: LegalService schema (schema.org now marks the older Attorney type as deprecated), clean crawlable pages. If an engine can't crawl and parse your site, it has less to go on. (Some sites also publish an llms.txt file, but Google's John Mueller said in June 2025 that no AI system was using it, so treat it as optional.)
• Reviews with substance: review text describes what you actually handled in a way a star rating can't. A review that says "handled my custody case" ties your firm to custody work in words any system reading it can see.

What GEO Work Actually Looks Like

In practice, a GEO engagement for a small firm is unglamorous: fixing entity data everywhere it appears, building out practice-area and location pages with real specificity, adding structured data, earning directory and citation consistency, and publishing the kind of answer content engines quote. Then measuring: asking the engines your target questions monthly and tracking whether, and how, your firm appears.

Why It Pays to Start Now

There is little published evidence on how quickly AI recommendations shift once a firm improves its signals, and the engines change their models and sources often. What the Citorian data does show is convergence: all ten of its most-named firms were named by all five engines, and those firms were established practices with heavy directory presence. Directory profiles, reviews, and citable content take time to build, so in our view firms that start building them sooner are better placed as the engines evolve.

We build SEO, AEO, and GEO into one website revamp for solo and small firms. See pricing. Start with the fundamentals: law firm SEO in the AI era.