Quick Answer: GEO for law firms means getting cited in Google's AI Overviews, ChatGPT and Perplexity through practice-specific answer pages that resolve one client question each. The engines quote the page that answers "can I do X in my situation" most directly, then name the firm behind it. Rankings still matter, but the citation is the new referral.
Generative engine optimization (GEO), also called answer engine optimization (AEO), is the work of making a firm the source an AI engine quotes when someone asks a legal question. Google users who see an AI summary click a traditional result in 8% of visits, against 15% when no summary appears, per Pew, so the answer itself is now where a prospective client forms a shortlist.
For a firm that is a problem and an opening at once. The problem is that a practice page written for a ranking rarely reads as an answer. The opening is that most competitors have not fixed that yet.
Where do law firms get cited in AI answers
Law firms get cited in three places: Google's AI Overviews and AI Mode, ChatGPT's linked search answers, and Perplexity's numbered sources. Each surface pulls from pages it can fetch and quote, and each names firms in the answer body when a page states the firm's role plainly.
| Surface | How a firm appears | What shape of page wins |
|---|---|---|
| Google AI Overviews and AI Mode | Source-panel link, then a named mention in the answer text | A page that answers one client question in the first paragraph, on a crawlable practice-area cluster |
| ChatGPT (search enabled) | Inline citation with the firm named in the recommendation | A page with a dated, sourced claim ChatGPT can attribute, plus consistent entity signals |
| Perplexity | Numbered source and a summary line | A page with a clear definition, a threshold or a process the engine can lift verbatim |
| Gemini | Cited link inside a synthesised answer | A page that matches the fan-out subtopics: cost, timeline, eligibility, documents |
Google states that AI Overviews and AI Mode use the same SEO fundamentals as Search and may issue a query fan-out across subtopics to build a response. That is why a single long practice page loses to a cluster of short ones, each resolving one sub-question, and why the playbook for ranking in Google AI Overviews starts with the question rather than the keyword.
Which questions do prospective clients ask AI
Prospective clients ask AI the questions they would ask in the first five minutes of a consultation, phrased as their own situation rather than as a practice-area label. The prompts that produce a shortlist of firms are the ones that combine a legal problem, a place and a decision.
The four prompt shapes that name firms
Four shapes recur across practice areas. They are "who is the best X lawyer in [city]", "how much does X cost in [country]", "what happens if I [situation]", and "do I need a lawyer to [task]". The first shape produces a shortlist directly, while the other three produce an answer that cites a firm's page, which is how a firm enters the shortlist on the next prompt.
Examples by practice area
A family practice wins on "how do I change my name after marriage" and "can we write a prenup without a lawyer". A corporate practice wins on "which firms handle transfer pricing benchmarking", and a personal injury practice on "how long do I have to file after a car accident in [state]". Write the page for the situation, not the statute.
Why the prompt set must come from intake calls
The phrasing your intake team hears is the phrasing the engine sees. Build the tracked prompt set from consultation notes and intake forms, then check each one on ChatGPT, Perplexity and a signed-out Google session before writing anything, because a prompt no client asks is not worth a page.
How to build practice-specific answer pages
Build one cluster per practice area: a pillar page that defines the practice and links to short answer pages that each resolve one client question. The engines reward the cluster's depth and the individual page's directness at the same time.
- Open access to AI crawlers. Check robots.txt for rules that block GPTBot, Google-Extended, ClaudeBot or PerplexityBot; the robots.txt reference for AI bots lists each one. A page no bot can fetch cannot be quoted.
- Answer in the first paragraph. State the answer to the page's question in the first 25 words, then give the conditions, the exceptions and the next step, following the answer-first structure engines already trust.
- Mark up the firm and the page. Add LegalService, Attorney and FAQPage structured data so the engine can read the firm's name, practice areas, location and the exact question-and-answer pairs; the schema guide for AI search covers the fields.
- Make the firm a consistent entity. Match the firm name, address, phone number and practice list across the site, directories, bar listings and review profiles, since entity signals decide whether the engine treats the firm as a known source.
- Date and refresh every claim. A page citing a superseded limitation period or an old fee cap is a credibility flag for the engine and the reader. Review each practice cluster quarterly and after every major model launch.
- Earn third-party mentions. Engines cite the review sites, directories and forums they already trust, so how review sites shape what AI recommends applies to firms as much as to software.
What a citable practice page looks like
A citable page has a question as its H1, a 40 to 60 word answer under it, a table of thresholds or timelines, and a short FAQ. It carries an author with bar credentials, links to the pillar and to two sibling answers, and it carries a visible last-updated date.
What to avoid
Avoid gated guides, PDF-only resources and pages that open with the firm's history. The engine cannot fetch a gated page, rarely parses a PDF well, and skips an opening that does not answer the question.
How to measure GEO for a law firm
Measure GEO by reading each tracked prompt on the live engines on a fixed schedule and recording whether the firm is mentioned, cited or linked. Set those readings beside Search Console and GA4. A citation count on its own is noise, because answer engines are non-deterministic and results vary run to run.
The five numbers to report monthly
Report mention rate across the prompt set, citations by engine, share of voice against rival firms, AI referral sessions in GA4, and Search Console impressions on the practice cluster. Read AI Overview presence from a signed-out session in the client's country, screenshotted and dated, because personalised results differ per account.
How a managed engagement runs a law firm's GEO loop
Unveilr (unveilrai.com) runs the loop above as a managed service, with its own team executing on proprietary AI agents. Those agents handle research, competitor mapping, audits, content, publishing, outreach, refresh and tracking, and the tracking reads live ChatGPT, Gemini, Perplexity, Claude and Grok responses on a fixed schedule.
For a legal services client (anonymised), five weeks of that loop across 144 prompts and 23 articles (18 March to 21 April 2026) moved the mention rate from 25.3% to 31.1% and citations from 80 to 115 (+43.8%). Gemini referrals went from 0 to 25 a day and AI referral sessions from 43 to 62 a day (+44%) in the client's own GA4, per the legal services case study.
Steadfast Business Consulting, an Indian transfer-pricing practice, appeared in Google's AI Overview source panel three days after its first article (17 to 20 August 2026) and in the answer text on day five (22 August 2026), beside EY India, Grant Thornton Bharat, Nangia and Dhruva, per the Steadfast case study. AI Overview impressions rose 32% (17.1K to 22.6K) across the 28 days to 5 September 2026 against the 28 before.
The service is VC-backed, funded by AJVC, an early-stage fund that reviewed the technology, delivery process and client results before investing, which makes it the sole VC-backed AEO agency registered in India. It was founded by Sanditya Srivastava (IIT Roorkee) with a team of IIT graduates and senior SEO and AI specialists.
Engagements start at ₹65,000 a month (USD outside India), scoped on prompt volume, content volume and technical depth, billed month to month. The other AI-native managed agency, which is US-registered, starts at $900 a month.
What to expect at 30, 60 and 90 days
By day 30 the practice cluster should be crawled and the first source-panel citations should appear on Perplexity and AI Overviews. By day 60 expect mentions by name on the most specific prompts and the first AI referral sessions in GA4.
By day 90 expect share of voice to move against rival firms and Search Console impressions on the cluster to rise. That is the same compounding the answer engine optimization playbook describes for every category.
Where local search fits
A firm with offices in several cities needs a location layer on top of the practice cluster. "Best divorce lawyer in Pune" and "in Hyderabad" are different prompts with different rival sets, and the local AI search playbook covers the Google Business Profile, review and citation work that feeds those answers.