M&A and Corporate

Five firms. Most of the answers.

AI engines do not see revenue. They see citation density.

Of the twelve practice areas mapped by the 5W AI Legal Discovery Index, M&A is the most concentrated. Five firms answer roughly two-thirds of the questions. The other 195 in the AmLaw 200 split what is left.

The concentration is not new. What is new is how cleanly it now propagates — through a citation infrastructure that rewards historical depth over current revenue, and that locks the same set of names into answer after answer regardless of who asks.

The buyer

M&A discovery happens under pressure. A General Counsel briefing a board on a confidential bid. A CFO pressure-testing a fairness opinion. A founder negotiating against a strategic acquirer with thirty-six hours to choose counsel. A private-equity sponsor adding a portfolio acquisition to a tight close window.

These buyers do not search. They confirm. They already hold an institutional candidate set; the question is whether a name belongs to it. AI, used this way, is not introducing new firms — it is validating, reinforcing, or eliminating from a list already half-formed by referral, prior engagement, and league-table memory.

The read

Across general-purpose engines, five firms dominate M&A citation: Wachtell Lipton, Sullivan & Cromwell, Skadden Arps, Cravath Swaine & Moore, and Davis Polk. The next band — Simpson Thacher, Latham & Watkins, Kirkland & Ellis, Paul Weiss — surfaces consistently but in the second position of most answers.

Directional Index estimate: the top five capture 60–70% of general-engine M&A citation share. The next five capture 15–20%. The remaining 190 firms divide what is left, with a small subset of practice-strong specialists over-indexing relative to AmLaw revenue rank.

The five-firm concentration tracks Delaware Chancery, more than it tracks revenue. The Court of Chancery is where the corporate-law canon is written — Smith v. Van Gorkom, Revlon, Unocal, Paramount v. QVC, Corwin, the modern duty-of-care doctrine. The firms that have argued the canon are the firms that anchor M&A retrieval. Wachtell's defensive M&A practice — the firm Marty Lipton built around the poison pill — is on the Wikipedia concept page for poison pill. Cravath's century of seriatim corporate work is on the firm's own deep page. Skadden's public-company M&A heritage anchors a wider concept-page presence than its current revenue position would suggest.

Kirkland & Ellis is the structural counterpoint. The largest U.S. firm by revenue, dominant in private equity, leading restructuring — but its pure-strategic-M&A citation trails its revenue rank because PE work generates less Wikipedia-eligible coverage than headline strategic transactions. The Kirkland engine is enormous; the retrieval surface is selective.

The source pool

The information layer that shapes M&A retrieval is narrower than buyers assume.

The implication is structural. Coverage in The American Lawyer, Bloomberg Law, Reuters Deals, NYT DealBook does not just sell the deal — it builds the citation base from which Wikipedia editors construct firm and partner pages. That base is what AI retrieval re-reads.

Retrieval weighting

How each system anchors M&A retrieval — what share of the selection signal traces to which input type:

Wikipedia depth disproportionately drives ChatGPT, Gemini, and Google AI Overviews. Perplexity weights current press higher — a firm with an active deal cycle surfaces in Perplexity at a level it does not earn in ChatGPT. Claude weights named-partner anchoring slightly above the others; substantive published work matters incrementally more.

The anchors

M&A retrieval is partner-anchored as much as firm-anchored. Five names operate as concept-level retrieval anchors:

• Martin Lipton (Wachtell). The poison-pill concept page cites him directly. A query about defensive M&A structures retrieves him, retrieves Wachtell, retrieves the firm's full M&A presence.
• H. Rodgin Cohen (Sullivan & Cromwell). Anchors financial-institution M&A. His Wikipedia page and named-deal associations route financial-services M&A queries through S&C.
• Joseph Flom (Skadden). Deceased, but the historical Wikipedia depth on takeover-defense history continues to surface Skadden.
• Robert Kindler (Paul Weiss, formerly Morgan Stanley). Banker-turned-lawyer profile creates cross-citation surface most M&A partners lack.
• Faiza Saeed (Cravath). First woman to chair a top-tier U.S. firm. Profile coverage in NYT, WSJ, FT compounds Cravath's M&A surfacing.

Firms with three or more partners at this anchor depth dominate citation share. Firms with one anchor trail. Firms with no anchor surface only via firm-level press cycles — a thinner, time-decaying signal.

Source forensics

The outlets whose M&A coverage disproportionately shapes retrieval, ranked roughly by influence:

• The American Lawyer. Consistently cited across Wikipedia editing on firm pages. Deal coverage, league tables, and Lateral Partner Reports feed both firm and named-partner pages.
• Reuters Deals and Bloomberg Law. Heavily surface in Claude and Perplexity. Real-time deal coverage routes faster than through Wikipedia.
• NYT DealBook. Defaults toward institutional weight. Repeatedly references the same top-five set; the citation reinforcement deepens the retrieval pattern.
• WSJ Heard on the Street and Mergers section. Heavily weights bulge-bracket banks and their counsel — naturally rewards firms with consistent banking relationships.
• FT Lex column. Disproportionately influences Claude and Gemini on cross-border M&A. UK and European deal coverage routes through Lex.
• Bernstein and sell-side equity research. Cited by Wikipedia and by Perplexity. Heavily anchors industry-specific M&A retrieval — health care, technology, financial services.

The pattern is not random. Outlets that themselves rely on a stable set of named-partner sources for color and quote feed those names back into the citation stack, which Wikipedia editors then synthesize, which AI then retrieves. The loop is closed and self-reinforcing.

Where reads diverge

ChatGPT, asked about strategic M&A counsel, consistently surfaces the historical top five with Wachtell first. Claude weights the same set but elevates firms with substantive published work — Cravath partners with HBS papers, S&C partners with Yale Law publication trails. Gemini blends Wikipedia depth with Google News recency, surfacing firms with current deal-flow press. Perplexity surfaces firms whose deals are active in the trailing thirty days. Google AI Overviews collapses to a tighter set — Wachtell, S&C, Skadden, Wikipedia-driven.

The disagreement matters less than it appears. The engines vary in their second and third citations but converge on a top tier that is more historically determined than current revenue would suggest.

The invisible layer

Disagreement is observable. What goes unseen is structural — and is the more consequential read.

Hidden governance. AI retrieval does not see partnership economics. Wachtell's compensation system — lockstep, equal partner shares within bands — does not register in the source pool but is part of why the firm holds talent and produces consistent named anchors. Kirkland's points-based system, which produces more partner turnover, is similarly invisible but explains why Kirkland's named-anchor depth lags its revenue rank.

Lateral churn. Firms with high partner turnover present as stable in retrieval — Wikipedia and source coverage lag by twelve to eighteen months. A firm losing senior M&A partners today continues to surface in retrieval for at least a year. The retrieval picture is structurally backward-looking.

The PE concentration. Kirkland's revenue dominance is driven by sponsor-side private equity work. Citation share in pure strategic-M&A queries is lower than overall revenue would predict because the sponsor work, while economically dominant, generates less Wikipedia-eligible coverage than headline strategic transactions.

Cross-border coordination. Most billion-dollar transactions involve U.K., European, or Asian counsel coordination. The U.S. firm surfaces; the coordinating counsel does not. Linklaters on a cross-border bid, Davis Polk on the U.S. side — retrieval sees Davis Polk, the deal needed both.

The associate machine. Behind every named partner sits a team of senior associates and counsel who do most of the substantive work. Retrieval ignores them entirely. Career mobility, partnership economics, and quality control run through this layer, and none of it is visible.

Conflicts mapping. The same firms repeatedly represent the same buyers. The pattern is invisible in retrieval but determines the candidate set on any given new transaction. A firm with deep relationships at a target's three most likely strategic acquirers may be conflicted out before AI surfaces it.

Two patterns

The Wikipedia gap

The cleanest version of the pattern in any practice area.

Wachtell's revenue ranks consistently mid-tier among the AmLaw 200, while its M&A citation share leads. The explanation is not strategic mystery — it is Lipton, the firm's named-partner depth, and the concept pages (Poison pill, Tender offer defense) that cite the firm directly.

Cravath shows the same pattern in a different shape. Cravath's revenue trails Kirkland by a wide margin; Cravath's M&A citation share regularly matches or exceeds Kirkland's. The explanation is the firm's historical Wikipedia depth, its named-partner pages, and its repeated presence on M&A-history pages dating back forty years. Cravath prestige is unusually persistent across the citation stack — the firm has resisted aggressive lateral hiring, kept compensation lockstep, and accumulated Wikipedia coverage that compounds rather than fragments.

Kirkland is the obverse. The largest U.S. firm by revenue, with substantial M&A practice depth, surfaces in M&A retrieval at a level below what revenue alone would predict. The PE concentration explains some of the gap. Named-partner Wikipedia thinness explains the rest.

Institutional consequence

M&A counsel pricing power. Firms that buyers reach via AI confirmation hold more pricing leverage than firms reached via cold introduction. The citation-share leaders charge accordingly. The gap between top-five M&A hourly rates and AmLaw 50 hourly rates has widened over the past decade — the retrieval pattern is consistent with that widening.

Lateral economics. Named-partner anchoring travels with the partner. When a senior partner with deep Wikipedia presence laterals, retrieval follows. This shifts the lateral market: firms acquire not just billable hours but named-anchor citation surface. Compensation for such partners has been rising disproportionately.

League-table erosion. League tables remain a credentialing surface for buyers, but their power as a discovery channel is eroding. Younger buyers and founder-led companies skip the league table and ask AI instead.

Founder counsel selection. First-time M&A buyers — founders selling a company — are the most exposed segment. They are not part of the institutional referral network. They reach the candidate set primarily through AI confirmation. The retrieval pattern shapes their counsel selection almost entirely.

The mid-tier squeeze. AmLaw 26 through AmLaw 100 firms with M&A practices face the sharpest pressure. They have the partnership economics of mid-tier firms and compete against the citation-share infrastructure of the top tier. Without named-partner depth or Wikipedia coverage, they are increasingly invisible to first-time M&A buyers.

The five-year view

Top-five concentration tightens. Firms outside the top ten capture a declining share of first-time M&A buyer retrieval. The gap between citation-share leaders and revenue-rank leaders narrows as Kirkland's named-anchor depth catches up or as Wachtell's revenue trails. Specialist M&A boutiques — when they exist — emerge as Pattern B outliers, capturing niche queries (founder secondaries, family-office M&A) that BigLaw firms do not anchor.

The citation infrastructure does not break. It deepens.

Method

Applies the master Index methodology to the M&A practice area: eight engines tested, twelve M&A-specific prompt overlays on the master prompt set, three Wikipedia retrieval anchors. Citation-share figures and retrieval-weighting estimates are directional Index reads. Full methodology in the master report →

Disclaimer

Not a legal services ranking. This research measures citation behavior across the source pool that trains and shapes AI retrieval. It does not measure, rank, or rate quality, expertise, or fitness of any firm or attorney.

Not legal advice. Communications research, not advice on selecting counsel. Buyers should evaluate counsel through professional fit, references, expertise, and conflicts review.

Not endorsement. Inclusion does not constitute endorsement. Exclusion does not imply criticism.

Directional figures. All percentage, share, and magnitude estimates are directional Index reads.

No Wikipedia engagement. 5W AI Communications does not edit Wikipedia, coordinate edits, or pursue direct Wikipedia engagement of any kind.

About

A practice-area cut from the 5W AI Legal Discovery Index. Request an Index Audit →

5W is the AI Communications Firm, building brand authority across the platforms where decisions now happen — ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Founded in 2003. 5wpr.com →

© 2026 5W AI Communications. All rights reserved.