How People Find a Lawyer in 2026: Search Is Slipping, AI Answers Are Rising, and the Middle Is Still Unmapped

The front door is moving, even if the house looks the same

Online research before ever contacting an attorney has gone from a majority behavior to close to a default one. Clio's 2025 figures put the share of legal clients who research attorneys online before making contact at 79%, up from 57% just three years earlier. That number reaches us through a secondary blog summary rather than Clio's primary report, which is why we're treating it as emerging rather than established, but the direction lines up with everything else in this evidence base: fewer people are walking into a law office cold, and more of the decision is being made before the phone ever rings. For a firm, that means the moment that used to happen in a consultation, the moment where a prospect forms a first impression of competence and trustworthiness, now happens earlier and somewhere else entirely: a search results page, a review profile, or increasingly, an answer an AI engine gives.

The moment a prospect forms a first impression now happens earlier, and somewhere else entirely.

Google still leads, but it lost ground faster this year than in any prior year measured

Google remains the single most-used source for both legal research specifically and local-business research generally. BrightLocal's 2025 survey found 83% of consumers use Google for local-business research, comfortably ahead of local news sites, Yelp, Facebook, YouTube, and TikTok. But within the legal category specifically, iLawyer Marketing's 2026 survey (n=1,110 US adults) recorded consumers who'd use Google to research a lawyer falling from 86.7% in 2025 to 71.9% in 2026, a 14.8-point single-year drop the researchers describe as the steepest in the survey's history. That's a proprietary, single-vendor figure, and we're tiering it emerging accordingly rather than treating it as settled fact. But the size of the move is hard to explain away as noise. Something changed in how people think about researching a lawyer between 2025 and 2026, and Google's share of that behavior absorbed most of the change.

The AI-chat habit is real, and it skews toward the buyers who actually need a lawyer

Willingness to use ChatGPT to research a lawyer climbed from 9% in 2023 to 21% in 2024, 28.1% in 2025, and 41.9% in 2026, per the same iLawyer survey. Just over half of consumers, 50.1%, said they'd use at least one AI-answer engine for attorney research, and 9.5% said they'd rely on AI sources exclusively. The detail that should reshape how a firm thinks about this shift is the age breakdown: 57% of consumers aged 45 to 60 said they'd use ChatGPT to research lawyers, compared with 35% of those aged 18 to 29. That inverts the assumption that AI-search adoption is a youth phenomenon. The 45-to-60 group is disproportionately the group with mortgages, businesses, custody arrangements, and estates, which is to say the group actually generating billable legal need. If this survey's direction holds, the buyers most worth reaching are also the ones moving fastest toward asking an engine before they ask a person.

Fewer clicks doesn't mean the click stopped mattering

Pew Research Center's July 2025 study, the most methodologically independent behavioral data in this evidence base, found that about 58% of Google users saw at least one AI Overview in a month of searching, though only around 18% of all searches actually triggered one. When a summary did appear, the click-through rate to a web link fell to 8%, versus 15% on a standard results page, and only 1% of users clicked a link inside the summary itself. This is general search behavior across all categories, not a legal-specific measurement, and no verified source in this evidence base tells us how often a legal query specifically triggers a summary in the first place. What Pew does add is a trust layer worth sitting with: among people who had seen an AI summary, 72% found it useful and 54% trusted the information at least somewhat, but only 6% of all Americans said they trust AI search summaries a lot. A prospect can walk away from an AI Overview with a favorable impression of a firm they never clicked through to, and mostly not fully trusting the source that gave it to them. That's a strange kind of visibility to compete for, and it's not one a click-through-rate report was built to measure.

A prospect can walk away with a favorable impression of a firm they never clicked through to.

Reviews are consolidating around fewer platforms, and the FTC has sharpened what a firm can do about them

BrightLocal's 2025 survey found 99% of consumers read online reviews before a purchase and 92% read local-business reviews before a first visit, spending an average of thirteen minutes and forty-five seconds reading roughly ten reviews before deciding to trust a local business. That makes review presence one of the most durable pieces of the discovery terrain, established rather than emerging, and it's also getting more concentrated: iLawyer's 2026 data shows Yelp usage for lawyer research falling from 24.3% to 13.6%, Reddit from 20.1% to 11.7%, and Facebook from 24.7% to 19.2%, all within a single year. At the same time, the regulatory ground under reviews shifted twice. The FTC's revised Endorsement Guides, effective July 26, 2023, made clear that a material connection must be disclosed whenever a significant minority of the audience wouldn't otherwise expect it, and that suppressing negative reviews while forwarding only favorable ones can itself be deceptive. The FTC's Trade Regulation Rule on Consumer Reviews and Testimonials, effective October 21, 2024, went further, banning the sale or purchase of fake reviews, review suppression, and fake indicators of social-media influence outright, with penalties up to $51,744 per violation. For a firm, the combination is unambiguous: reviews carry more weight with buyers than any single directory, and the federal government is now actively enforcing how a firm is allowed to collect and manage them.

State bars loosened the advertising leash in the same window regulators tightened it elsewhere

While the FTC narrowed what firms can do with reviews, several state bars widened what firms can say in their own advertising. The ABA's modernized Model Rules 7.1-7.3 are rolling out state by state through 2026: Alabama's version took effect January 1, 2026, now permitting lawyers to list contact information instead of a physical office address in ads, and permitting lawyers to describe themselves as "specialists" based on experience, education, or training rather than requiring formal certification, provided the claim isn't false or misleading under Rule 7.1. California's Senate Bill 37, signed in 2025 and effective January 1, 2026, made comparable revisions to attorney-advertising regulation under the State Bar Act. The throughline across both is permission with a leash still attached: firms have genuinely more latitude in how they describe their practice and reach prospective clients, but the underlying honesty standard, no false or misleading claims, hasn't moved at all. More freedom to say more, in other words, is not freedom to say anything.

Paid local capture still works, but the economics deserve a hedge, not a headline

Two figures round out the paid side of the terrain, and both come from industry trade sources rather than an audited primary study, which is why we're holding them at contested rather than established. Google reportedly discontinued its separate "Google Guaranteed" and "Google Screened" trust badges for Local Services Ads in late 2025, unifying them into a single "Google Verified" badge, itself a small signal that even the platform side of this terrain is still being redesigned. On cost, trade sources put legal Local Services Ads leads at roughly $150 to $300 depending on practice area and market in 2026, with personal injury leads converting to retained clients at around 34%, which implies a true cost per signed case closer to $630 to $735. Directionally plausible, worth planning against, not something to repeat to a client as an audited number.

The justice gap sets a ceiling on how much channel-shifting can move actual demand

None of this discovery-channel migration changes how much legal need exists in the first place, and the Legal Services Corporation's 2022 study is a sobering check on that. Low-income Americans received no or inadequate legal help for 92% of the substantial civil legal problems they experienced. Seventy-four percent of low-income households had at least one civil legal problem in the prior year, yet only 25% of those with a substantially impactful problem sought any legal help at all, and 46% of non-seekers cited cost as the barrier. Channel-shifting, whether toward Google, a directory, a review site, or an AI engine, only affects how existing, resourced demand finds its way to a firm. It doesn't create legal need that isn't there, and it can't fix the affordability barrier that keeps most substantial civil legal problems from ever reaching a lawyer at all. Any read of this evidence that implies AI search is growing the market overall is overreaching what the data shows.

Channel-shifting only affects how existing demand finds its way to a firm. It doesn't create legal need that isn't there.

Where this leaves a firm: the unmapped middle

The summary of this evidence base is that the edges are well measured and the middle is not. We know, from Pew, how a general search user behaves around an AI summary once one appears. We know, from BrightLocal, how much weight reviews carry once a prospect starts reading them. We know, from the FTC and two state bars, exactly what a firm is and isn't allowed to say across all of it. What we don't know, because no independently verifiable source currently answers it, is how often a legal question specifically produces a synthesized answer in the first place, which engines are answering it, and what those answers say about any given firm. A widely circulated claim putting that trigger rate at 78% for legal queries, the highest of any industry vertical, traces only to a public-relations release citing an unpublished analysis, and could not be verified at the primary source, so we're not repeating it as fact. That gap is exactly the kind of thing that has to be measured directly, market by market and firm by firm, rather than assumed from a press release.