Where AI Finds Lawyers: The Sources That Earn Citations.

Why off-site sources decide it

An AI engine does not trust a single website, no matter how good. It trusts agreement across sources.

When you ask an engine for a lawyer, it has to decide which firm is safe to name. A polished firm site is one voice making a claim. The engine grows confident when several independent, authoritative sources, a directory profile, a bar record, a news mention, a community thread, describe the same firm the same way. That corroboration is what turns a claim into evidence, and it happens almost entirely off the firm's own site.

This is why off-site presence matters as much as on-page work, a pattern we measured directly in our five-engine study of which firms AI recommends. The firms the engines named were the ones with a deep, consistent footprint across the sources below.

Where AI actually looks (the data)

We logged the sources the engines cited across the answers from the four engines that expose their links: ChatGPT, Perplexity, Claude, and Google AI Overviews. Gemini hides its sources behind redirect links, so it is excluded from this tally. A handful of sources did most of the work.

Two things stand out. Super Lawyers and Justia are close to mandatory, cited far more than any other source. And Reddit outranked the older lawyer directories Avvo and Yelp, with YouTube outranking Yelp too, a sign the engines weigh community discussion and video alongside the traditional directories. The full method is in the study itself.

The directories that do the heavy lifting

If two sources are close to mandatory for AI citation, they are the major legal directories, led by Super Lawyers and Justia.

The engines lean on directories because they are structured, independent, and built to verify lawyers, and two kinds matter here. The claimable profiles, Justia, Avvo, FindLaw, Martindale-Hubbell, and Lawyers.com, where the work is to claim each listing, complete it fully, and make sure the firm name, attorney names, practice focus, location, and bar admissions read identically to your own site. And the recognition-based listings, Super Lawyers and Best Lawyers, with its Best Law Firms ranking, where a place is earned through peer selection rather than claimed, so the work is to qualify and then keep the resulting profile accurate. Both kinds showed up heavily in the data, Super Lawyers and Justia most of all.

Completeness and consistency matter more than sheer presence. A half-filled or stale profile, or one where the firm name or focus differs from your site, gives the engine a reason to doubt rather than corroborate. Treat each profile as a verification of the same facts, not a separate marketing page.

Bar and court records

The most trustworthy corroboration is the kind a firm cannot write itself: official records.

State bar profiles, attorney license and admission records, and court and case records are among the highest-trust sources an engine can reach, because they are independent and verifiable. Make sure your attorneys' bar profiles are current and that the names and credentials on your site match the official records exactly. Where a firm has appointed roles, recognitions, or results that appear in public records, those become corroboration the engines weigh heavily, especially on the high-stakes questions where trust matters most.

Reddit and the community surprise

The clearest surprise in the data was how often the engines cited Reddit, ahead of the older lawyer directories.

AI engines increasingly treat community discussion as a signal of real-world reputation. People ask for lawyer recommendations in local and practice-specific forums, and the engines read those threads. You cannot, and must not, manufacture this.

Astroturfing violates platform rules and bar advertising rules, and engines and moderators are good at spotting it. What you can do is earn it: be genuinely helpful where your potential clients gather, answer questions without pitching, and build the kind of reputation people mention on their own.

Be careful here. Anything that looks like a planted endorsement or an undisclosed solicitation crosses both the platform's rules and the bar rules on advertising and solicitation. The goal is to be worth mentioning, not to mention yourself.

YouTube and earned media

Video and earned press round out the sources, and both reward genuine substance.

YouTube was cited more than Yelp in our data, which reflects how often the engines pull from educational video. Clear, substantive videos that answer the questions clients actually ask, what to do after an accident, how a specific claim works, build a presence the engines can cite. Earned media and trade press work the same way: a quoted comment, a bylined piece, or a recognition is an independent source describing the firm's expertise. None of this is quick, but it compounds, and it is exactly the kind of third-party authority a directory profile alone cannot provide.

Your own site: necessary, not sufficient

Your website does not earn the citation, but it makes the citation possible.

The engines still need to read your site to understand who you are, so it has to be clear, accurate, answer-shaped, and marked up with complete structured data so the facts are unambiguous. But the site is the assertion; the sources around it are the proof. A firm with a beautiful site and no off-site footprint reads as one unverified voice.

The two work together: the site states the facts, and the directories, records, and mentions confirm them. The on-page half of this is covered in the per-vertical guides, such as how PI firms get cited and how med mal firms get cited.

The signal that ties it together

The single thing that turns scattered listings into corroboration is consistency.

An engine connects sources into one trustworthy firm when the details match: the exact firm name, the attorney names and spellings, the practice focus, the location, the bar numbers. Inconsistencies, an old firm name on one directory, a different specialty on another, a name spelled two ways, do the opposite, giving the engine reason to treat the listings as separate or unreliable. The highest-leverage early project for most firms is not adding new sources, it is making the ones they already have say exactly the same thing. This is the same name, address, and phone consistency discipline that wins the local map pack, covered in the Map Pack playbook.

Three ways firms get the sources wrong

Three mistakes account for most of the gap between a firm's website and its actual AI visibility.

Where to start

You do not need to be everywhere at once. You need to be consistent where it counts, in the order the data suggests.

Most firms find the same thing: their website is strong and their off-site footprint is thin and inconsistent, and closing that gap is the fastest path to being the firm the engines name. To see your own footprint first, our free AI visibility check shows where you currently stand across the engines.