When someone asks an AI assistant a legal question, the answer usually names two to five sources. Getting your firm into that short list is the new front page. Citation is earned differently than a ranking: the engine needs a page it can quote cleanly, an entity it can verify, and permission to read your site in the first place. This guide walks through all three, in the order that actually matters.
Start with the mechanics: how AI answers pick their sources
Every major answer engine — ChatGPT with search, Google's AI Overviews and AI Mode, Gemini, Perplexity — works on the same loop: interpret the question, retrieve a set of candidate pages, then synthesize an answer while attributing specific claims to specific pages. The retrieval step leans heavily on classic search signals, which is why conventional SEO still matters. But the attribution step is different: the model quotes pages that contain clean, self-contained, verifiable statements. A page that buries its answer in the fourth paragraph of a story rarely gets quoted, even when it ranks.
That means citability is a writing problem as much as a technical one. The pages that win citations state the answer in the first sentence of a section, attach concrete numbers or jurisdictions to it, and leave no ambiguity about who is making the claim.
Step 1 — Make sure AI crawlers can read you at all
Before content strategy, check the plumbing. OpenAI's GPTBot and OAI-SearchBot, Google-Extended, PerplexityBot, and Anthropic's ClaudeBot each respect robots.txt. If your firm's site blocks them — and many law firm templates shipped with blanket bot-blocking during the 2023 scraping panic — you are invisible to the engines you want citations from. Allow the crawlers you want, and keep your substantive content server-rendered: most AI crawlers do not execute JavaScript, so a React page that assembles its text in the browser reads as empty.
- Allow GPTBot, OAI-SearchBot, Google-Extended, PerplexityBot, and ClaudeBot in robots.txt
- Serve real HTML — practice-area content must exist in the page source, not only after JavaScript runs
- Add an llms.txt file that lists your key pages with one-line descriptions
- Keep canonical URLs clean and consistent so engines consolidate signals on one address per page
Step 2 — Write statements an engine can lift
Audit your top pages with one question: could a machine quote a single sentence from this section and have it stand alone as true? "Our experienced team fights for you" cannot be cited. "In Ontario, the basic limitation period for most civil claims is two years from discovery" can. Rewrite section openings so the answer comes first and the explanation follows. Add jurisdiction, dates, and dollar figures wherever the law actually supplies them — and keep them current, because a stale number is worse than none. AI engines increasingly cross-check claims across sources, and a page that disagrees with the government source loses.
FAQ sections earn outsized citations because they mirror the question-answer shape of the AI interaction itself. Give every important page five to eight real questions with direct two-to-four sentence answers, marked up with FAQPage schema.
Step 3 — Give the engine an entity it can verify
Engines cite sources they can identify. A law firm that exists as a verified entity — consistent name, one canonical website, Organization and Attorney schema, a Google Business Profile, professional directory listings that agree with each other, and named lawyers with Person schema and real bios — is a safer thing to cite than an anonymous blog. This is the entity SEO layer, and it compounds: every consistent external reference makes the next citation more likely.
Named authorship matters more each year. Articles bylined by a real lawyer, connected by schema to a bio page with credentials and bar admissions, outperform anonymous content for citation because the engine can attribute expertise to a person, not just a domain.
The mistakes that keep good firms uncited
Most firms that "do SEO" and still never get mentioned are failing in one of four predictable ways. They publish content with no facts in it — a thousand words of empathy and zero quotable statements. They scatter one article across every topic instead of building depth in a few, so no engine sees them as the authority on anything. They let their entity footprint contradict itself — an old suite number in one directory, a merged-away partner name in another — which makes the engine hedge by citing someone cleaner. Or they gate everything behind JavaScript, chat widgets, and interstitials that crawlers cannot get past.
The common thread is that none of these are penalties. The engine is not punishing the firm; it simply found a safer source to quote. Fix the specific failure and the citations arrive without any further persuasion — which is also why buying links or stuffing keywords does nothing here. There is no trick to being quotable; there is only being the cleanest verifiable answer available.
Step 4 — Measure it like a channel
Once a month, run your ten most valuable client questions through ChatGPT, Gemini, Perplexity, and Google with AI Overviews. Record whether your firm is mentioned, which page was cited, and who was cited instead. Referral traffic from chatgpt.com and perplexity.ai now shows in analytics, so tag it and watch its consultation rate — AI-referred visitors arrive pre-educated and convert at notably higher rates than cold search traffic. Treat the engines that never cite you as a gap analysis: the sources they prefer show you exactly what your pages are missing.
Citation is not a hack; it is what happens when a verifiable expert publishes quotable answers that machines are allowed to read. Do those four things consistently and mentions follow.
Frequently Asked Questions
Grow your AI SEO practice with AI
Lexscale.ai builds AI search visibility, websites, and intake systems for ai seo firms across North America. Book a free strategy call to see what would move the needle for your practice.
Book a Free Strategy Call →