Method note: this article uses published survey data and named case studies only, all linked below. Survey sources that sell related services are labeled as vendor research. We make no claim in this article that we cannot trace to a linked source.

The first conversation between a potential client and a law firm increasingly happens without the firm in the room. The client describes the problem to a chatbot and gets back a short list of names. llmlisted's 2026 survey (871 US adults, Q2 2026) found 62% have used AI to look for a lawyer and 74% have used it to check a firm's reputation. llmlisted sells AI visibility services to law firms, so read the numbers as directional rather than exact. Even discounted, they describe a client acquisition channel that did not exist three years ago.

Most firms have not adjusted. The investment still goes to Google rankings while the entry point moves to answers that never show a results page.

What AI answers read when they recommend a firm

Answer engines do not keep a directory of law firms. They compose a recommendation from what has been written about firms: ranking lists, bar and directory entries, press coverage, review pages. The citation research that exists for other industries (Parse's study of 5.17 million cited links) consistently shows a minority of citations pointing at the vendor's own site. Answers repeat the written record, and for most firms that record is thin beyond their own website.

The compression problem

Ask several engines "who are the top personal injury firms in [city]" and the same well-known names come back. That is not a conspiracy, it is repetition: answers lean on lists that already exist, and the lists mostly cover big firms. A boutique with strong results but no published record sits outside the pool of candidates. No dataset currently quantifies this gap by firm size, so we state it as the mechanism the citation studies support, not as a measured number.

What winning looks like: Kingwell Immigration Law

A Toronto immigration firm, Kingwell, worked with the agency WEBRIS on local search, review signals, and content aimed at high-intent questions. Within months, inbound leads and calls tripled, and the firm hired additional attorneys to handle the volume. The campaign predates AI answers and won on Google. We include it because the assets that produced the result (specific practice-area pages, review volume, local trust signals) are the same written record an AI answer reads when it composes a recommendation. Firms that own that record for their niche are findable in both eras.

The niche advantage

Cordell & Cordell, the largest US family law firm by its own description, writes its entire blog for one reader: fathers going through divorce. The titles are the reader's own questions. That specificity is also how people phrase questions to AI. A firm that has published thorough answers to the exact questions its niche asks has handed every engine raw material no generalist competitor has.

A compliance note

Legal marketing is regulated, and AI adds a twist: an engine can describe your firm wrongly (wrong practice areas, wrong jurisdiction, outdated fee models) and potential clients will take that description at face value. CiteGap's scope is diagnosis only. We test how AI engines describe your firm across real client questions and report what is missing or wrong. We do not write legal content for firms.

Questions your potential clients are already asking

Illustrative examples, phrased the way clients phrase legal searches:

  • "Do I need a lawyer to contest a non-compete agreement in California?"
  • "Who are the top personal injury firms near me?"
  • "What should I look for when hiring an estate planning attorney?"

Run your list against ChatGPT, Perplexity, and Gemini and record which firms come back and how they are described. Or start with CiteGap's free scan (5 questions, 1 engine). For law-firm-specific details, see our AI visibility guide for law firms.

Get the written diagnosis ($490)

Frequently asked questions

Do clients really use AI to find lawyers?

llmlisted's 2026 survey (871 US adults, Q2 2026) found 62% have used AI to look for a lawyer and 74% have used it to check a firm's reputation. The source sells AI visibility services to law firms, so treat the figures as directional. The direction itself is not in dispute: the research is consistent that client research has partially moved into AI tools.

Why doesn't my firm appear in AI answers?

Most firms have a thin written record outside their own site: few directory entries beyond the basics, no ranking-list presence, little press. Answers compose from that record and repeat what already exists, which is why the same large firms come back for most queries. A firm can change the record, but only by publishing material the engines can quote.

Is this just SEO with a new name?

The underlying assets overlap: specific practice-area pages, reviews, directories, and press all feed both channels. The question is different. SEO ranks a page; an AI answer names firms, and your firm is either in the named list or absent from it. Measuring rank tells you nothing about the second outcome, which is why the diagnostic has to ask the questions directly.

Can a boutique firm beat the big names in AI answers?

In its own niche, plausibly. The big names dominate general queries because the existing lists are about them. A boutique that publishes the most thorough public answers to its niche's exact questions gives engines raw material the generalists lack. No result is guaranteed, which is why the first step is diagnosis: find out what the engines currently say about your firm.

Sources