Google AI Overviews are erasing law firm SEO. What comes next.
The page-one playbook that built every legal marketing agency for 25 years just stopped working. Three things every firm needs to do in the next 90 days — none of them are "more blog posts."
For 25 years, law firm marketing was a single, well-understood game: rank on Google for "personal injury lawyer + city," collect the click, convert the call. Around that one mechanic an entire industry of legal SEO agencies built itself. As of late 2025, that mechanic broke — and most firms haven't yet realized how badly.
The break has a specific cause: Google AI Overviews now answer the question above the organic results, and AI-native search engines like ChatGPT, Perplexity, and Claude don't show ten blue links at all. When a prospect types "do I need a lawyer for a DUI in Onondaga County" or "what's the deadline to file a slip-and-fall claim in California," the AI gives them a synthesized answer. The lawyer they retain is the one cited inside that answer, not the one ranked third on Google's old SERP.
That changes everything about how a law firm has to think about visibility — and the firms that figure it out in the next two quarters are going to compound, while the firms that keep paying SEO agencies for "page one rankings" on terms that no longer have a page one are going to quietly lose share.
What actually changed.
A few things broke at once and the cumulative effect is worse than each piece individually:
- AI Overviews push organic results below the fold. Even when you rank #1, the click-through rate has fallen sharply because the answer is now sitting at the top of the page in Google's AI summary.
- Discovery is fragmenting. A meaningful share of legal questions — by some estimates 25–35% — now route through ChatGPT, Perplexity, Gemini, or Claude before ever touching Google. Those platforms cite a different set of sources than Google's organic results.
- Local has held up better than national. "Personal injury lawyer Syracuse" still pulls a map pack and local results. But broader informational queries — the top of the funnel — have largely moved to AI answers.
- The bar for citation has gone up, not down. AI engines tend to cite content with strong topical authority, structured information, named author expertise, and verifiable facts. Thin SEO content is being filtered out, not promoted.
Why most legal SEO agencies are giving the wrong advice.
The dominant playbook from legal marketing agencies right now is some version of "produce more content, target more keywords, build more backlinks." That advice is roughly the same advice that's been given for fifteen years, with the words "AI" and "GEO" sprinkled in. It misses the actual mechanic.
The mechanic that matters in 2026 isn't ranking. It's citation. Whether your firm gets named when an AI engine answers a prospect's question. That's a structurally different optimization target than "get to position 3 on Google." It rewards different things — depth over volume, expertise over keyword density, authority signals AI models actually weigh over generic backlink counts.
The 2010s asked: do you rank? The 2020s ask: does AI know you exist?
The three things every firm should be doing in the next 90 days.
1. Audit which AI engines already cite you — and don't.
This is unbelievably simple and almost no firm does it. Open ChatGPT, Perplexity, Claude, and Google's AI Overview. Type the ten most important informational queries in your practice area and geo. See which firms get cited. If yours doesn't, you have an immediate, specific gap to fix. If your competitors do, you have a roadmap. We do this in the first week of every diagnostic; the gap between what firms think their AI visibility looks like and what it actually looks like is usually startling.
2. Restructure your highest-traffic content for AI citation.
AI engines tend to cite content that has clear, structured answers — definitive statements, real data, statutory references, jurisdiction-specific specifics, named expertise. A 2,500-word "ultimate guide" that hedges every claim and reads like it was written by committee gets ignored. A 600-word page that gives the actual answer, cites the actual statute, and is signed by an actual attorney with credentials gets cited.
The work, practically, is to take your 20–30 most important practice-area pages and rewrite them around answer-engine optimization principles: structured headings, clear factual claims, statutory citations where relevant, named author expertise, and updated dates. This is the single highest-ROI content effort available to law firms right now, and it cannibalizes most of what the SEO agency wants to bill you for.
3. Build authority signals that AI engines weight, not just Google.
Backlinks still matter, but the lazy "bought-link" strategy that legal SEO agencies have run for years is actively counterproductive now. The signals AI engines weigh more heavily are: citations from trusted publications, presence in legal directories that the model trusts, authored expertise (real bylines on real platforms), and consistent factual presence across the web. That means PR, original research, real op-eds, statutory commentary, and showing up where actual humans write about legal topics — not "we'll get you 50 backlinks from law-firm-blog-network-7."
The firms that win the next decade are going to look more like publishers than like marketers. The agencies that cling to the old playbook are going to keep selling rankings on terms that no longer drive business.
What this means for marketing budgets.
A specific recommendation, because abstract advice is annoying: in most firms we audit right now, we recommend reallocating roughly 20–30% of the existing SEO budget away from keyword-volume content production and toward (a) answer-engine optimization on existing top pages, (b) authentic authority-building — original commentary, real bylines, real PR — and (c) measurement infrastructure that actually tracks AI-engine visibility, not just Google rank. The rest of the budget often stays, but pointed at smarter targets.
For paid media, the shift is less dramatic but still real. Local Services Ads and well-targeted Performance Max are holding up. National PI campaigns running on broad-match Google Ads against AI-summarized queries are not. The intelligence to figure out which campaigns to keep and which to kill requires looking at attribution that ties spend to signed cases — which most firms still don't have, and which is half of what we install in the operations engine.
The opportunity, briefly.
Most law firms are going to do nothing about this until 2027 or 2028, by which point the firms that moved in 2026 will have compounded an authority advantage that's expensive to displace. The firms that move first don't need to be radical or risky about it — they just need to stop running a 2019 playbook in a 2026 environment, and reallocate the same budget toward what actually works now.
That's the opportunity. It's specific, it's quantifiable, and it has a 90-day window where the cost of action is low and the cost of inaction is structural.
If you want to see what AI engines say about your firm — and what your competitors look like in the same lens — that's literally the first deliverable of the diagnostic. Start there →