AI visibility for law firms - what works?

Q: What actually moves AI visibility for law firms?

Directory depth on Avvo, Justia, FindLaw, Martindale-Hubbell, Super Lawyers and Best Lawyers; LegalService and Attorney JSON-LD schema; substantive Google reviews; case results pages within ABA Model Rule 7.1 limits; and practice-area pages with state-specific procedural depth. Blog volume and ChatGPT-specific 'AI SEO' packages do almost nothing.

Q: Does writing more blog content help a law firm in AI search?

Generic high-volume blogging does not. The engines already have many versions of common explainers indexed. State-specific, statute-specific, or unusually detailed treatments earn citations; thin general posts are noise.

Q: Which legal directories carry the most weight in AI answers?

Avvo and Justia tend to dominate the citation pattern across engines, with FindLaw and Martindale-Hubbell close behind. Super Lawyers and Best Lawyers profiles also surface, especially for higher-end practices. State bar listings come up mostly on disclaimer-heavy queries.

Q: How long does it take to move a firm's AI visibility?

Weeks, not days. Engines re-crawl directories on their own cadence and ChatGPT mixes training-data lag with live browsing inconsistently. Real movement shows up over weeks of consistent directory hygiene, schema, and content depth.

Question

AI visibility for law firms - what works?

Maciej Grabek · Accepted Answer

For law firms, AI visibility is won on the same surfaces that have always carried legal authority on the open web: Avvo, Justia, FindLaw, Martindale-Hubbell, Super Lawyers, Best Lawyers, and the firm's own substantive content. ChatGPT, Perplexity, Gemini, Claude, and DeepSeek read these sources. Blog volume does almost nothing. Directory depth, schema, real case results, and review substance do almost everything. The shortlist of what actually works We have run AI-visibility audits on dozens of firms across personal injury, family, criminal, immigration, and estate planning. The pattern is consistent enough to summarise. Firms that get cited share most of these traits: Complete profiles on the legal directory stack. Avvo, Justia, FindLaw, Martindale-Hubbell, Lawyers.com, Super Lawyers, Best Lawyers, and the state bar listing. "Complete" means photos, full bios, peer endorsements, primary practice area set correctly, NAP matching the firm site to the suite number. LegalService or Attorney JSON-LD schema on the homepage and every attorney bio page. Address, geo, openingHours, areaServed, sameAs links to the directory profiles, knowsAbout listing the practice areas. Schema does not rank you. It removes ambiguity for Google AI Overviews and Gemini about which entity the page describes. Substantive Google reviews that mention the practice area and the city by name. Star count alone is not enough. AI engines lift the language from review text when summarising firm reputation, so "helped me with my custody case in Plano" is worth more than three five-star reviews with no body text. Case results pages with real outcomes, within the bounds of ABA Model Rule 7.1 and your state's analogue. Verdicts, settlements (where disclosable), motions won, charges dismissed. Engines lift these in answers about firm credibility. Practice-area pages with depth, not 400-word stubs. A page on "contesting a will in [state]" that walks through statutory deadlines, grounds for contest, and procedure outranks ten thin pages titled "estate litigation services". What does not work, despite what vendors will sell you Three categories of effort move the needle far less than vendors claim: High-volume blog content. Publishing two posts a week of generic "what to do after a car accident" content is largely noise. The engines already have ten thousand versions of that page indexed. They prefer state-specific, statute-specific, or unusually detailed treatments. Schema for the sake of schema. Stuffing FAQPage and HowTo schema onto every page can trigger Google quality signals against the site. Use schema where it describes the entity (LegalService, Attorney, BreadcrumbList, Article on substantive posts). Skip the rest. Generic "AI SEO" packages that promise to optimise for ChatGPT specifically. There is no ChatGPT-specific tactic in 2026. ChatGPT with browsing reads live search results plus its own training. The same fundamentals that earn citations in Perplexity and Gemini earn them in ChatGPT. A real prompt to test against Try this verbatim across ChatGPT, Perplexity, and Gemini: "Best [practice area] lawyer in [city], with case results." Note which firms appear, which directories are cited, and how often Avvo or Justia profile pages show up versus firm sites. The directories tend to dominate the citation pattern across engines. Firms that own the citation are the ones whose own site has the depth to compete with a directory profile - case results, statute-level practice content, and real attorney bios. What to do this quarter if AI visibility matters to the firm Audit current visibility across all five engines for the firm name, the top three practice-area-plus-city queries, and the top two competitors. This is what we run as the entry point of the engagement. Reconcile NAP across Avvo, Justia, FindLaw, Martindale-Hubbell, Lawyers.com, Super Lawyers, and the state bar. One mismatched suite number is enough to suppress a profile. Add LegalService and Attorney JSON-LD on the firm site. Validate it. Skip schemas that do not describe real entities on the page. Identify the top three practice-area pages on the firm site and rewrite them with state-specific procedural depth. Replace generic stub content. Set up a monthly tracker for AI engine mentions so the change can be measured against a baseline, not eyeballed. The honest reality on timeline AI visibility for legal does not move in a week. The engines re-crawl directories on their own cadence, and ChatGPT in particular reflects training-data lag plus live-browsing freshness inconsistently. In our experience, real movement on engine citation patterns shows up over weeks of consistent directory hygiene and content depth, not days. Firms expecting a 30-day turnaround are setting themselves up to fire whoever they hired.