By The LexText Team
Updated:
Jul 27, 2026
Elite partners are building smaller, technology-enabled practices. The lesson is not that every firm must start over. It is that litigation workflows must change.
Introduction
David Fox helped devise the growth playbook that transformed Kirkland & Ellis. Now he is betting on a very different model.
As The Wall Street Journal recently reported, Fox co-founded Irving Technology, the company powering a new law firm called Irving. The firm has fewer than 10 lawyers and engineers. Its technology handles work such as research and drafting so senior lawyers can focus on judgment, client relationships, and the hardest parts of a matter.
Irving is focused on deals. But the same shift is reaching litigation.
In March, Benjamin Gruenstein left Cravath after 14 years as partner to launch a boutique firm specializing in investigations, white-collar defense, regulatory enforcement, and complex civil litigation. He described a model built around “small teams of exceptional lawyers” augmented by transformative legal technology.
In July, Chris Kercher left Quinn Emanuel after more than 17 years to launch Kercher Law. Kercher had founded Quinn Emanuel’s AI and Data Analytics Group and spent years applying AI to commercial litigation. He told Bloomberg Law that starting a new firm allowed him to reconsider the roles of partners, associates, paralegals, software, and AI from first principles.
These departures don't signal the disappearance of traditional firms. But they do suggest that experienced partners believe smaller, technology-enabled teams can deliver high-stakes litigation work differently.
That should get every litigation leader’s attention.
The real innovation is the operating model
The most visible feature of an AI-native firm is its size. The more important feature is how it organizes the work.
Traditional firms often add technology to an existing process. The same people perform the same sequence of tasks, with a new tool available at certain points.
An AI-native firm begins with a different question:
If powerful AI already existed, how would we staff the matter, organize the work, train lawyers, communicate with the client, and price the result?
Established firms should not discard the institutional knowledge, talent, and client relationships that make them valuable. But they can apply the same first-principles thinking to how legal work gets produced.
The immediate goal is not to become an AI-native law firm. It is to build AI-native litigation workflows inside the existing firm.
Litigation raises the bar
Litigation is adversarial, fact-intensive, and constantly changing. Relevant information is scattered across pleadings, discovery, correspondence, depositions, expert reports, and case law. Each new filing or piece of evidence can shift the strength of a claim, defense, or argument.
A general-purpose chatbot can produce a polished summary or plausible draft. But polished prose is not litigation analysis.
Consider a new complaint. Before drafting an answer, a litigator must:
Identify the claims and their required elements
Map the material allegations to those elements
Surface potential defenses and counterclaims
Identify factual and legal gaps
Anticipate the opposing side’s strongest response
Decide what the case is really about
The answer is only one output of that reasoning.
The same is true throughout a case. Effective discovery follows from the disputed issues. A useful deposition outline connects the pleadings, documents, and prior testimony. A strong motion begins with an argument structure and command of the record, not a request for several pages of persuasive prose.
Legal AI should support that reasoning, not skip over it.
Four principles for AI-native litigation workflows
Clients are signaling what they expect
Across professional services, generative AI adoption is accelerating. The 2026 Thomson Reuters AI in Professional Services Report found that 40% of professionals say their organizations now use generative AI, up from 22% the year before.
Two-thirds of corporate respondents want their outside firms to use AI. Yet only 18% of organizations track its return on investment.
The firms that stand out will not be those that announce the most tools or report the most logins. They will be those that can show how AI improves legal work: faster command of the case, stronger analysis, fewer avoidable revisions, more senior attention, and better use of the client’s budget.
Where established firms should begin
A firmwide transformation program is not required.
Start with one recurring, consequential workflow, such as complaint analysis, discovery, deposition preparation, motion practice, or document search.
Then:
Define what excellent work looks like.
Establish the current time and review burden.
Set the security and verification requirements.
Run the workflow on real matters.
Measure both time and quality.
Expand based on evidence rather than enthusiasm.
This captures much of the advantage of an AI-native model while preserving what already makes the firm valuable.
Bringing litigation-native AI to the existing firm
LexText was built by litigators around this approach.
Rather than treating litigation as a series of isolated prompts, LexText provides structured workflows for case analysis, legal research, document search, pleadings, discovery, motions, deposition work, and other litigation tasks. It helps lawyers surface stronger arguments, work from matter materials, and produce litigation-ready work product while retaining control over the result.
LexText is SOC 2 Type II, never trains on customer data, and includes safeguards built for the accuracy and confidentiality demands of litigation.
AI-native does not have to describe the law firm. It can describe how the firm’s best lawyers work.
See LexText on your real work. Bring a recent motion, complaint, or discovery request for a private 30-minute session with a litigator on our team.