Roughly 78% of legal search queries now trigger a Google AI Overview. That’s the highest rate of any industry, ahead of healthcare, finance, and retail. The result sits above paid ads and pushes traditional organic listings down the page, often below the fold.
Pew Research tracked 68,879 Google queries in 2025. When an AI Overview appeared, users clicked through to a traditional result 8% of the time. Without one, they clicked 15% of the time. That’s a 47% relative decline in click-through rate.
For a law firm, this isn’t a ranking problem. It’s a visibility problem. You can rank number one and still lose half your traffic.
What Changed in the Search Result
The AI Overview answers the user’s question directly. Often it cites two to four sources. Sometimes it cites none. The user reads the summary and either clicks a cited source or closes the tab.
The numbers compound. AI Overviews don’t just suppress clicks. They terminate sessions. A quarter of the time, the search ends right there. The user never visits any law firm site, never compares options, never picks up the phone.
The firms losing the most traffic are the ones that built their entire acquisition strategy on ranking. The firms losing the least are the ones already known by name in their market.
Cited Beats Ranked
Here’s the inversion. Sites that get cited inside the AI Overview see brand clicks rise, even as overall organic clicks fall. The traffic shrinks but qualifies harder. AI-referred visitors convert at roughly 3x the rate of standard organic traffic.
Citation isn’t a function of keyword density. It’s a function of trust signals. The model has to decide which sources to surface in a 200-word summary. It looks for entity authority, original data, and consistent presence across the web.
The boring 2,000-word “what to do after a car accident in [city]” page doesn’t get cited. A firm with a real brand, real reviews, real citations from authoritative legal directories, and original data published on its own domain does.
This is the core of generative engine optimization. The metric is no longer rank. The metric is whether the model trusts you enough to name you in the answer.
Why Legal Got Hit Harder Than Any Other Vertical
Three reasons.
First, legal queries are mostly informational. “Statute of limitations.” “What does no fault mean.” “How long does a settlement take.” The AI Overview can answer these directly without sending the user anywhere. Compare that to ecommerce, where the user needs to actually buy a product on a specific site.
Second, legal SERPs were already saturated with thin, templated content. Forty firms in the same metro publishing the same “we handle these case types” copy. The AI doesn’t need any of them. It synthesizes a generic answer that captures what they all said and shows nobody.
Third, legal content has high authority bars. Wikipedia, Cornell Law, government sites, and a handful of established legal publishers dominate the citations. A regional PI firm rarely makes the cut for general informational queries. The firms that do make the cut are usually those with original case data, named expert authors, and consistent third-party references.
What Gets Cited
Sites that earn AI Overview citations in legal queries share four traits.
A defined brand entity. Consistent name, address, attorney bios, credentials, and Schema.org markup across the firm site, Google Business Profile, legal directories, and news mentions. The model needs to confirm the entity is real. A firm that exists as eight different name variants across the web won’t get cited.
Original data or research. A statistic, study, or breakdown that doesn’t exist anywhere else. The AI prefers primary sources. A firm that publishes its own annual local case data, its own settlement breakdowns, or its own market research becomes a primary source. A firm that aggregates other people’s data does not.
Structured schema markup. FAQ schema, HowTo schema, Article schema, Person schema for attorneys, LegalService schema for the firm. The model parses these to confirm what the page is about and who’s behind it. Sites without schema get parsed by guesswork.
High-authority citations. Mentions in legal directories with editorial standards (Super Lawyers, Best Lawyers, Avvo with verified credentials), bar association profiles, news coverage in mainstream outlets, and academic citations. These off-page signals tell the model the entity is trusted in the broader ecosystem.
None of this is keyword-driven. The era of optimizing a page for “best personal injury lawyer in [city]” by repeating the phrase 17 times is over. The new optimization is at the entity layer.
What Doesn’t Get Cited
A pattern emerges from the sites that consistently miss citation across legal queries.
Templated practice area pages with the same structure as 50 other firms. AI-generated blog content without original analysis. Sites with thin Person schema or no attorney bios. Firms with inconsistent NAP (name, address, phone) data across the web. Pages without any structured markup at all. Content that summarizes Wikipedia instead of adding to the conversation.
The most painful pattern: firms that paid an SEO agency $5K to $15K per month for years to publish thin content. That investment produced a content library that the AI categorically ignores. The work didn’t fail. The system changed underneath it.
The Real Cost of the Click Decline
A personal injury firm generating 200 organic leads per month from its website at a cost per signed case of $2,400 was doing the math at $480K in cost per month for organic-attributed cases. Cut that organic traffic by 47% and you’re now generating 106 leads at the same fixed SEO investment. The cost per signed case nearly doubles overnight.
The firms that diversified months ago aren’t feeling this. The firms that didn’t are watching their entire business model bend.
The March 2026 core update compounded the problem. Sites already weakened by AI Overview suppression got further demoted by the algorithm shift toward entity-driven ranking. The firms running content farms lost 20-35% of organic traffic in March. The firms with real brand entities mostly held.
How Firms Get Cited Now
The work splits into four categories. None of it is fast. All of it compounds.
Build the entity. Standardize the firm name across every web presence. Update Google Business Profile, legal directory listings, social profiles, news bylines. Add full Person schema for every attorney with credentials, education, and bar admissions. Add LegalService schema for the firm with practice areas, service area, and aggregate review data. The goal is for the model to recognize one entity, not three variants.
Publish original data. A firm with proprietary case data, local market research, or unique analytical perspective becomes citation-worthy. We publish legal advertising spend data across 210 markets because nobody else does. A PI firm could publish its own settlement data, recovery rates by claim type, average case timelines, or local accident data. Real numbers from a real practice.
Earn third-party citations. Pitch local news outlets on commentary about legal news. Publish bylined analysis in legal industry publications. Get listed in editorially curated directories. Submit to Super Lawyers, Best Lawyers, and Chambers nominations. The off-page mentions are what tell the model the entity is trusted.
Structure the on-page signal. Implement FAQ, HowTo, Article, and LegalService schema across the site. Use clean heading hierarchies. Write content in question-and-answer formats that match how queries are phrased. Make the content easy for an AI to extract and summarize.
The output of all this work is not a ranking. It’s a citation rate. Track which queries cite the firm in the AI Overview. Track which queries cite competitors. Track which queries cite no one. The competitive map looks different than the keyword ranking map.
The Channel Google Can’t Touch
Even firms that nail GEO will see total search traffic shrink. The AI Overview is a permanent compression of organic search demand. The way around it is to generate demand outside of search.
CTV advertising reaches the same 35-64 demographic that files personal injury claims. Completion rates run 90-96% because the spots are non-skippable. CPMs sit at $12-22 in most markets. The firms running CTV brand campaigns generate branded search demand that AI Overviews can’t suppress, because branded search bypasses the Overview entirely.
When someone types your firm name directly into Google, no AI Overview interrupts. The user sees the firm’s site, the local pack, and the reviews. The conversion path stays clean.
This is why brand investment now beats content investment. A 30-second CTV spot that runs in your DMA produces measurable branded search lift within weeks. A 2,000-word blog post produces nothing the AI Overview won’t immediately summarize and bury.
What to Do This Quarter
The firms that come through 2026 with their acquisition pipelines intact will share a profile. Strong brand entity. Original data published consistently. Clean schema across the site. Diversified channel mix that doesn’t depend on organic search clicks.
Audit the entity first. Pull every web mention of the firm and check for consistency. Fix the schema next. Most legal sites have nothing structured beyond basic Article markup. Then start producing original data. Quarterly local market reports, annual settlement summaries, original perspective on legal news in the firm’s practice areas.
While that work compounds, build a brand channel that doesn’t go through Google. CTV in the firm’s primary DMA. Targeted broadcast in mid-size markets. Direct mail in high-value zip codes. Anything that produces branded search demand and household-level recall.
The firms still calling SEO their growth strategy in late 2026 will be losing market share to the firms that figured out the entity game two years earlier. The cited firms get the AI traffic. The brand firms get the direct traffic. Everyone else gets compressed.
The window to build citation authority is open right now. It won’t be open in 18 months.
References
- Pew Research Center. "Google Users Are Less Likely to Click on Links When an AI Summary Appears in the Results." July 2025.
- Search Engine Land. "Google's AI Overviews are hurting clicks, Pew study finds." July 2025.
- SE Ranking. "AI Overviews Industry Analysis: Legal Queries Trigger Rate." 2026.
- Previsible. "AI Referral Traffic Study: 19 GA4 Properties Tracked January-May 2025." 2025.
- Search Engine Journal. "Pew Research Confirms Google AI Overviews Is Eroding Web Ecosystem." 2025.
