A few years ago, someone looking for a divorce attorney or a DUI lawyer would open Google, scroll through ten blue links, and start calling firms. That habit hasn’t disappeared, but it’s no longer the whole story. A growing share of people now open ChatGPT, Perplexity, or Google’s AI Overviews and simply ask: “Who’s a good personal injury lawyer in my city?” or “I got rear-ended yesterday, do I need a lawyer?”

The assistant answers in a few sentences. Sometimes it names two or three firms. Sometimes it just explains what to look for. Either way, that answer shapes the person’s shortlist before they ever land on a law firm’s website. For firms, this raises a question that simply didn’t exist five years ago: when someone asks an AI for a lawyer, does your firm come up — or does a competitor?

This isn’t a theoretical shift. It’s measurable, and the data on it is more specific than most marketing advice admits.

AI Doesn’t Rank Pages. It Picks Sources to Trust.

Traditional search engines rank pages. AI assistants do something different — they read across many sources, decide which ones are credible, and synthesize an answer. There’s no page-one, page-two. There’s “cited” or “not cited.”

That distinction matters because it changes what actually earns a mention. A law firm can rank well on Google and still get skipped entirely by ChatGPT, simply because the assistant couldn’t verify the firm’s details confidently enough to include it. Conversely, a firm with a modest website but a complete, consistent presence across a few trusted sources can outperform a much bigger competitor in AI answers.

A firm becomes “visible” to these systems when its name, location, and practice areas show up consistently and are corroborated by outside sources — directories, reviews, local coverage, professional listings. A firm becomes invisible not because it’s being penalized, but because the AI simply doesn’t have enough corroborated information to act on. Faced with that gap, the model plays it safe: it either stays vague (“look for someone with experience in this area of law”) or names whichever competitor it can describe with more confidence.

What the Data Actually Shows

Several independent studies published in 2026 have looked specifically at what AI engines cite when answering legal questions, and the numbers are worth knowing rather than guessing at.

How often does a firm even get named?

A study by the AI-visibility research firm Citorian, which logged 359 answers across five major AI engines in five U.S. metro areas, found that a specific firm was named in 62% of answers overall. But the phrasing of the question mattered enormously. Direct, commercial-style prompts like “best personal injury lawyer in [city]” produced a named firm 94% of the time. Situational prompts like “I was in a car accident, who do I call” produced a name only 18% of the time — the model tended to explain the process instead of recommending a firm.

Two other findings from that study are worth sitting with. First, the five engines largely agreed with each other: every one of the ten most frequently named firms in the study was named by all five engines tested. That suggests AI visibility isn’t five separate battles for five different platforms — it’s closer to one contest, because the engines are drawing on overlapping sources. Second, advertising spend didn’t appear to move the needle. One of the largest personal injury advertisers in the country showed up in only a handful of the logged answers, which fits the broader pattern: these systems are reading independent, third-party content, not media budgets.

Which sources do the engines actually cite for legal queries?

The same Citorian study found a fairly concentrated picture: Super Lawyers accounted for roughly a third of the source citations behind legal AI answers, Justia around a fifth, Google listings about 12%, Reddit 8%, and Avvo only around 5%. A separate 2026 industry report from 5WPR and Haute Lawyer, which tested legal categories across ChatGPT, Claude, Perplexity, and Google’s AI Mode, reached a similar conclusion from a different angle: a small group of about seven directories — Super Lawyers, Justia, Avvo, Martindale, Chambers, Legal 500, and Best Lawyers — accounted for nearly all citations across the legal categories it tested, with the last three mattering more for business and corporate law searches than for personal injury or family law.

How does this compare to the wider web?

A separate, cross-industry analysis of over 680 million AI citations (compiled by the analytics firm Profound and cited in 5WPR’s broader 2026 report on AI citation patterns) found that Wikipedia alone made up nearly half of ChatGPT’s most-cited sources, while Reddit accounted for close to half of Perplexity’s. Only about 11% of the domains cited by ChatGPT were also cited by Perplexity for the same queries — a reminder that these platforms don’t necessarily read from the same well, and firms shouldn’t assume that ranking with one guarantees anything with another.

Does traditional Google ranking still matter?

Somewhat, but less directly than it used to. An analysis by Ahrefs of roughly 4 million URLs cited in Google’s AI Overviews found that only about 38% of cited pages also ranked in the traditional top 10 for that search — down sharply from a similar analysis run less than a year earlier. Almost a third of cited pages weren’t even ranking in the top 100. Google itself, in its public documentation on AI features, states plainly that there’s no special markup or extra requirement to appear in AI Overviews — a page simply needs to be indexed, eligible to show a normal search snippet, and genuinely answer the underlying question.

Why This Happens: The Engines Reward Specificity, Not Polish

A widely cited academic study out of Princeton (the “GEO” paper by Aggarwal and colleagues, presented at KDD 2024) tested what actually made web content more likely to be pulled into generative AI answers across roughly ten thousand test queries. The clearest lever wasn’t backlinks or design — it was specificity. Pages that included concrete statistics, direct quotations, and clear source citations saw their odds of being surfaced rise by somewhere between 22% and 41%, depending on the topic.

For a law firm, the practical translation is blunt: a page that says “How much does a DUI cost in Ohio?” and then actually answers with real fee ranges and real timelines will get pulled into AI answers far more often than a page that talks around the question for six paragraphs before offering a “call us to find out” ending. The engines are looking for a passage that directly and confidently answers a specific sub-question — not a page optimized to keep a human scrolling.

Freshness plays a role too. Pages cited by AI tools tend to skew noticeably newer than the pages that rank highest in ordinary Google search, which is one more reason a visibly dated “last updated” line and an annual content refresh aren’t just cosmetic.

What This Actually Looks Like in Practice

Putting the research together, a few concrete priorities stand out for a firm that wants to show up when someone asks an AI assistant for legal help.

Get the basic facts consistent everywhere.

AI systems build their picture of a firm by cross-checking many sources — the firm’s own site, its Google Business Profile, its state bar listing, its directory profiles. If the firm’s name, address, phone number, or listed practice areas don’t match exactly across those, the system sees several weak, unverifiable entities instead of one strong one. This is unglamorous work, but it’s foundational: pick one canonical version of everything down to how the street address is abbreviated, and correct every listing to match it.

Prioritize the directories that are actually cited.

Based on the studies above, a complete, detailed profile on Super Lawyers and Justia does more for AI visibility than almost anything else a consumer-facing firm can do. Avvo and a strong Google Business Profile matter next. Firms serving businesses or handling larger litigation should add Martindale, Chambers, Legal 500, and Best Lawyers to that list, since those sources carry more weight for corporate-style queries. It’s worth noting that Super Lawyers’ own published selection process states that self-nominations aren’t accepted and that inclusion isn’t for sale — so the leverage there is making sure an existing, complete profile reflects real peer recognition, not buying placement.

Publish pages that actually answer questions.

Rather than general “why choose us” pages, content that answers one specific question thoroughly — with real numbers, real timelines, and real process steps — is far more likely to get pulled into an AI-generated answer, and it also tends to perform better in ordinary search.

Show up authentically where the community already talks.

Reddit’s outsized share of citations, especially on Perplexity, means that a named attorney genuinely answering questions in a local legal subreddit — with disclosed affiliation and no sales pitch — builds exactly the kind of independent corroboration these systems look for. Faked or astroturfed participation tends to get noticed and backfires; the value here comes specifically from it being real.

Don’t assume one platform speaks for all of them.

ChatGPT currently leans heavily on reference-style sources like Wikipedia and directories. Perplexity leans on community discussion and fresher content. Google’s AI Overviews pull from the broader indexed web, independent of normal ranking position, and are fed directly by Google Business Profile data for local queries. A firm that only optimizes for one of these is, in effect, only competing for a third of the audience.

Check monthly, not once.

Source mixes shift quickly — one industry study found Reddit’s share of ChatGPT’s citations dropped from roughly 60% to roughly 10% in the span of a single month in late 2025, a swing isolated to that one platform. A simple monthly habit — running the same handful of prompts across ChatGPT, Perplexity, and Google, and logging whether the firm is named and which sources get cited — turns this from a one-time project into something a firm can actually track and adjust.

The Bottom Line

Being findable by AI isn’t about a trick or a plugin. It comes down to the same thing trustworthiness has always come down to: consistent, verifiable facts about the firm, corroborated by independent sources the AI already trusts, paired with content that answers real client questions directly instead of talking around them. Firms that treat this as ordinary due diligence — cleaning up their listings, filling out their directory profiles honestly, and publishing genuinely useful answers — are the ones AI systems currently have enough to work with. Everyone else is, for now, simply invisible by default, not by penalty.

Related Reading

• SEO Content Optimization: A Practical Playbook for 2026 — a broader look at building content that performs well for both traditional search and AI-driven answers, from the team at Team Disquantified, whose SEO and IT services cover the technical groundwork (indexability, site structure, content strategy) that underpins AI visibility work like this.