Ex-Quinn Emanuel partner Christopher Kercher on building an AI-native litigation boutique — and why BigLaw's partner-flight problem is about to accelerate

Ex-Quinn Emanuel partner Christopher Kercher on building an AI-native litigation boutique

Christopher Kercher spent 17 years at Quinn Emanuel — representing clients including Citadel in the GameStop saga and Elon Musk in the Twitter litigation — before leaving in July 2026 to found Kercher Law, a five-lawyer litigation boutique built around AI from day one. The pitch is not cost savings. It's leverage.

The firm model

Kercher's core argument is that AI lets a small litigation team do what previously required armies of associates, without the coordination drag that slows large firms down. At five lawyers, Kercher Law keeps decision loops tight — particularly important in litigation, where coordination failures are costly. Kercher frames the technology not as a research assistant but as "super intelligence" applied to an inherently intelligence-driven problem: figuring out what matters in a case before the other side does. John Quinn's line, which Kercher cites directly, captures it — the side that figures out what matters first usually wins.

The practical application he describes is building persistent, structured case knowledge: client objectives, constraints, risks, all loaded into a system that can surface relevant context in real time. AI reminds the team that a witness is testifying in another case next week, or that an earn-out is coming up and the client may not want to escalate. The coordination that used to require layers of associates becomes a knowledge architecture problem.

“I'm 47. When AI blew up, like, I've seen this movie before. And I think I can take my now experience and expertise and use what I've learned about AI to build a new kind of firm. We don't need as many people... The whole game in litigation is to get stuff in your context window, in your head. So you can't really outsource the whole thing to AI. AI has to be used to get in your brain so that when you're standing at the podium arguing with the judge or cross examining a witness, you can call what you have to.”

Why BigLaw's structure breaks down

The Weil Gotshal dissolution frames the second half of the conversation. Kercher's read is that BigLaw has the revenue-concentration dynamics of a bank without deposit insurance. Revenue sits with a small number of rainmakers; when they leave, it goes with them. The shift away from lockstep compensation toward guaranteed lateral contracts has eroded the institutional loyalty that historically kept partners in place during stress. When one rainmaker exits, others reassess — the same run dynamic that takes down banks.

Kercher expects the first quarter of 2027 to be the next destabilizing moment. Most BigLaw partners collect their annual comp in January, which means departures cluster in Q1. The free-agency market in litigation is already active, and the economics of leaving have improved. A partner who keeps 80% of their book but drops the overhead of a large firm can come out ahead. The need for hundreds of associates — the main historical reason to stay — weakens when AI compresses the work that junior lawyers used to do.

His structural prediction is a barbell: a cluster of smaller, more agile litigation boutiques and potentially one or two dominant large survivors, with the mid-tier most exposed. Conflict-of-interest constraints accelerate this — large firms generate conflicts that force clients to shop elsewhere, and a small boutique with no such entanglements can compete directly for the same matters.

The limits AI doesn't cross

Kercher is specific about what AI cannot replace. Courtroom advocacy and client counseling are his two examples. The analogy he uses is a medical one: Claude can diagnose a cough, but a worried patient still sees a doctor. A Stanford law student who had never litigated could technically ask an LLM the same questions Kercher asks — but would have no framework for evaluating the answers, no instinct for the pitfalls, and no sense of what the model gets wrong.

The courtroom constraint is harder. Trial is compressed and irreversible. A month of trial means nothing else exists — no associates, no agents, no real-time AI feed changes the fact that the lawyer standing at the podium needs everything already in their head. AI is useful for getting material into the lawyer's context window before that moment. It cannot substitute for the moment itself.

Kercher Law is currently running a California roadshow, with meetings in LA and Stanford, as Kercher builds out the firm's client base from its Greenwich and New York base.