Part of a series. See the hub: Buyers Ask ChatGPT Before They Hire a Lawyer.
The top 50 firms in the AmLaw 100 dominate the legal directories. They lead the Chambers rankings. They run nine-figure marketing departments. And inside ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, they are losing the answer to firms a tenth their size.
This is the Citation Share gap. Run the prompt "best M&A law firms for cross-border deals" inside the major AI engines and Wachtell Lipton, Skadden Arps, and Sullivan & Cromwell get named — but so does a tier of mid-market specialists with stronger digital footprints, more recent third-party citations, and cleaner schema. Run "law firms that handle SEC parallel proceedings" and Davis Polk and Paul Weiss appear less consistently than firms that have published original analysis on Wells notices in the past 90 days. Brand authority offline does not translate to citation density online.
The Citation Share gap
Citation Share — the percentage of AI-generated answers in which a brand is named across a defined buyer-intent prompt set — exposes a hard pattern in legal. The largest firms consistently underperform their brand strength inside the engines. Kirkland & Ellis, the largest revenue-generating firm in the country, surfaces less often than its market position would predict when buyers query private equity practice areas. Latham & Watkins, a top-three global firm, gets out-cited in capital markets queries by mid-market firms with sharper content programs. Sullivan & Cromwell, a defining M&A name for a century, often does not appear in the top three answers to "best M&A firms in 2026."
The gap is not a measurement artifact. It is structural. The engines weight third-party citation density, schema-clean entity data, primary-source authority, and recency of mention. Big Law has the brand. It has not built the footprint.
Why brand size doesn't translate
A century-old name builds reputation through relationships, referrals, and rankings. The AmLaw 100 marketing infrastructure was built around those inputs — Chambers submissions, conference panels, Am Law profiles, client alerts. None of those inputs were designed to be retrieved by a machine.
Three structural failures recur across the largest firms. First, the firm website is a brochure. Attorney pages list bar admissions and case captions but omit the structured entity data the engines need to identify who practices what. Practice pages describe service offerings without naming the deals, regulators, or industries that anchor authority. Second, the firm's third-party citation density is concentrated inside legal trades — Law360, The American Lawyer, Bloomberg Law — and rarely extends into business press, industry publications, or research outlets the engines also weight. Third, the firm publishes client alerts as marketing content. The engines do not cite client alerts. They cite analysis with named authors, dated entries, sourced claims, and entity-rich context.
Where the largest firms are getting out-cited
The firms winning Citation Share against the AmLaw top tier share a recognizable profile. They are smaller. They are specialists. They publish dated, sourced, primary-source analysis on narrow questions. They have built schema-clean attorney pages. They get cited inside non-trade publications because they pitch outside the legal trades.
In commercial litigation, Susman Godfrey and Quinn Emanuel out-cite several AmLaw 25 firms inside ChatGPT and Perplexity for plaintiff-side commercial queries. In appellate work, Munger, Tolles & Olson and Williams & Connolly appear in answer sets where larger firms with full appellate practices do not. In white collar, boutiques with named partner brands — built around former Assistant U.S. Attorneys and SEC enforcement directors — surface in answers about specific investigative postures (parallel proceedings, declination strategy, monitorship management) where AmLaw 50 firms with 80-lawyer white collar groups do not.
The pattern is consistent: the engines reward specificity, recency, and structured authority. Firms that publish a 3,000-word original analysis of a single enforcement action with named attorneys quoted as the authority get cited. Firms that publish a thirty-page general-counsel guide do not.
The website is no longer the destination
The largest firms still treat their websites as marketing destinations. The buyer no longer reaches them. The buyer reads the AI answer and either contacts a firm or does not. The website matters now because the engines crawl it to identify who the firm's attorneys are, what they practice, and how they are credentialed. A schema-clean attorney page — degree, bar admission, case data, publication markup, named expertise — populates the engine's answer. A brochure attorney page does not.
The fix is technical, not creative. Generative Engine Optimization (GEO) rebuilds attorney and practice pages as structured entity graphs. The firm's authority becomes legible to the engines for the first time. Most AmLaw 100 firms have not done this work. The firms that move first will set the citation patterns the engines learn from.
What this means for laterals and clients
Partners considering lateral moves are now running AI Citation Share checks alongside the firm visit. A partner with a recognizable name inside ChatGPT and Perplexity for her practice area carries portable authority a firm cannot manufacture. Firms with high Citation Share are easier targets for lateral recruitment because the engine validates the firm's positioning before the candidate walks in. Firms with low Citation Share in their stated specialty face a recruiting headwind that no Chambers ranking solves.
Clients are running the same check, in reverse. A general counsel comparing three firms for a regulatory matter now runs the engines first. The shortlist the engine generates is the shortlist the GC trusts. Firms that get named win the call. Firms that do not, do not.
The work ahead for Big Law
The path is specific. Audit current Citation Share across the firm's core buyer-intent prompts on the five major engines. Identify the practice areas where the firm is losing the answer. Rebuild attorney pages and practice pages as structured entity graphs. Replace client alerts with dated, sourced, primary-source analysis under named bylines. Expand third-party citation beyond the legal trades. Measure monthly. The firms that begin this work in 2026 will set the patterns. The firms that wait until 2027 will spend the next five years trying to dislodge competitors who got cited first. See the companion piece, How Law Firms Win the AI Answer, for the five-layer GEO stack.
Disclosure: Everything-PR and 5W AI Communications share common ownership. Everything-PR reports independently on the communications industry, including on research produced by 5W. Editorial decisions are made by Everything-PR's editorial team.
