Which Personal Injury Lawyers Does AI Actually Recommend?

Why we ran this

Everyone in legal marketing now talks about getting cited by AI engines, but almost no one has measured what the engines actually do.

A growing number of people ask ChatGPT, Perplexity, Claude, Gemini, or Google AI Overviews for a lawyer before they ever type a query into a traditional search box. The firms those engines name capture the demand. The firms they skip never enter the conversation. We wanted real numbers on who gets named, how often, and why, instead of the usual assertions. So we ran a fixed, repeatable test and recorded every answer.

How we ran it

The method is deliberately simple and repeatable, so anyone can check it or rerun it.

We selected five US metros (New York, Los Angeles, Chicago, Houston, and Phoenix) and, for each, ran five buyer-intent queries for personal injury: a direct ask, a situational ask phrased the way an accident victim would, a specific-injury ask, a vetting ask, and a question-led ask. We ran every query across ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews, in logged-out sessions to limit personalization, three times each to capture variance. That produced 359 logged answers (from 375 query runs; a few sessions returned errors or no answer), collected between June 13 and 15, 2026. For each answer we recorded which firms were named, in what order, whether the engine declined to name anyone, and what sources it cited.

The headline numbers

Across 359 logged answers, the engines named at least one specific firm 62 percent of the time. But that average hides the real story, which is that the wording of the question decided almost everything. Ask in a clearly commercial way and a firm gets named the large majority of the time. Ask the way a person actually does in the first frightened minutes after a crash, and the engines name almost no one.

Why phrasing decides everything

The same engine will name a firm or refuse to, depending only on how the question is phrased. This was the clearest pattern in the whole study.

The split is not subtle. When the question signals commercial intent, that the person is choosing a lawyer, the engines name firms freely, between 79 and 94 percent of the time. When the question reads as an emergency or a request for guidance, they switch into a different mode.

On "I was in a car accident, who do I call," the engines lead with call 911, seek medical attention, document the scene, and contact your insurer, and they name a specific firm in only 18 percent of those answers. The firm recommendation, if it comes at all, comes later, or only when the person asks again more directly.

The safety-first response is intuitive. What is not obvious is the size of that gap, or what it means for where a firm should compete.

For a firm, the lesson is precise. The visibility worth winning sits behind the commercial-intent questions, the direct, vetting, and specific-injury asks, where the engines are actually willing to name someone. Optimizing to be the answer to "who do I call right now" is largely optimizing for an answer the engines decline to give.

How often each engine names anyone

The engines differ sharply in how willing they are to name a specific firm at all. Across all query types, ChatGPT, Perplexity, and Claude each named a firm about 60 percent of the time. Gemini named them most readily at 69 percent, though on a smaller, partly manual sample, and Google AI Overviews was the most reluctant, declining or giving only general advice on 41 percent of queries.

The most-cited firms

A small number of firms accounted for a large share of all the names the engines gave. The recurring names were established firms with strong legal-directory profiles, a mix of elite trial practices and high-volume firms. Across the 359 answers the engines produced more than 1,600 firm mentions spanning over 500 distinct firms, yet a core of roughly a dozen names came up far more often than any others. Notably, Morgan and Morgan, the largest national personal injury advertiser, was named in only a handful of answers, the kind of gap between ad spend and AI visibility we cover in beating the billboard giants.

The most striking part is the agreement between engines. All ten of the most-named firms were named by all five engines. The engines are not each cultivating their own favorites; they are converging on the same shortlist, which means visibility is a single shared contest rather than five separate ones.

What the engines lean on

Where an engine gets its names tells you exactly where a firm needs to be present. Across the four engines that expose real source links, a handful of directories did most of the work, with one platform far ahead of the rest.

Super Lawyers and Justia were cited far more than any other source, so directory presence is close to a prerequisite for being named. Two surprises stand out: Reddit outranked both Avvo and Yelp, and YouTube outranked Yelp, which suggests these engines weigh community discussion and video alongside the older lawyer directories. One caveat: Gemini returns opaque redirect links rather than real source URLs, so its underlying sources could not be attributed and are excluded from this tally.

What the cited firms had in common

This is the part that matters for any firm that wants to be named: the traits the most-cited firms shared.

We did not score each firm's website against a checklist for this first edition; that audit is the next step. But the pattern in the names themselves is already clear, and it lines up with the source data.

• They are established, named-partner firms. Panish Shea, Clifford Law, Gair Gair, Block O'Toole, Greene Broillet, the Lanier firm: every name at the top of the list is a senior, recognized practice built around named attorneys, not an anonymous brand.
• They have heavy legal-directory presence. The firms the engines named are the same firms that populate Super Lawyers and Justia, the two sources the engines leaned on most. Directory presence and being named travel together.
• Ad spend alone does not decide it. The list mixes heavy advertisers such as Lerner and Rowe and Phillips Law Group with trial firms that advertise comparatively little, while Morgan and Morgan, the largest national advertiser, was named in only a handful of answers.

The takeaway for a firm that wants to be named is consistent with everything else in the study. Be a named, credentialed practice with real depth, and be present and consistent across the directories the engines trust, above all Super Lawyers and Justia. That combination, not ad spend, is what the cited firms share.

Limitations

This is a snapshot, not a verdict. AI engines vary between sessions and change their models and sources often, so the exact figures will move; we ran each query more than once and dated everything to be transparent about that. The sample is five metros and one practice area, chosen for focus rather than completeness, so the numbers describe this slice rather than the whole country. We controlled for personalization by using logged-out sessions, but we cannot rule out all of it. None of this measures the quality of any firm. Being named by an engine is a visibility signal, not an endorsement, and nothing here is legal advice or a recommendation of any particular firm.

Run this for your own firm

You do not need our spreadsheet to see where you stand. You need to ask the engines the way your clients do.

Wherever your firm is missing and a competitor is named, you have found a gap worth closing, and the traits above tell you what tends to close it.