Build Your Own AI: Law Firms Betting on Proprietary Platforms

· LawDiver News Tracker

A living digest of law firms building proprietary AI systems — Morgan & Morgan MX2, Kirkland, Goodwin, BakerHostetler, Cleary, and more.

Build Your Own AI: Law Firms Betting on Proprietary Platforms

This post tracks a specific pattern in legal technology: law firms that build (or acquire to build) their own AI systems, rather than simply licensing a third-party legal AI workspace. It is a living digest. Every claim below is drawn from published reporting and firm announcements; when new coverage appears, this page is rewritten to incorporate it.

Scope note. This tracker deliberately excludes vendor platforms sold into law firms as primary products. The focus is firm-owned builds, firm-led co-builds with infrastructure partners, and acquisitions used to stand up internal product teams.

Why “build your own” is the story

Contingency and efficiency-receptive practices have a structural incentive to automate work they cannot bill by the hour. Artificial Lawyer’s coverage of Morgan & Morgan’s MX2 announcement framed plaintiff firms as especially “efficiency receptive”: AI that lowers operating cost and increases throughput maps cleanly onto a no-win, no-fee model (Artificial Lawyer, 14 Sept 2026). Big Law’s proprietary builds are different in motive — institutional knowledge, client confidentiality, and practice-specific workflows — but the pattern is the same: own the stack, or at least the layer that sits on top of foundation models.

Morgan & Morgan and MX2

According to Artificial Lawyer, Morgan & Morgan announced it will spend at least $1 billion on legal tech and AI over the next decade and formally unveiled MX2, its own legal AI platform. The same piece notes the firm had already spent roughly $300 million on technology over the prior five years, and situates the new commitment against Reuters’ June 2026 report of about $2.4 billion in firm revenue (Artificial Lawyer; Reuters on stake-sale exploration, 5 June 2026).

Artificial Lawyer describes MX2’s published capability set as including:

  • MX2 Agent — case insights, trial prep, and agentic workflows embedded in Litify
  • First Draft — case-aware document generation with Microsoft Office integration
  • Medical — medical-record extraction, chronology building, and alerts
  • Search — semantic search across unstructured case data

The firm’s own announcement of its first Chief AI Officer, Kaushik Chakraborti (formerly of Amazon), emphasizes that Morgan & Morgan has been “building proprietary technology that standardizes and scales firm operations,” with a tech leadership bench spanning transformation, CTO, CIO, product, and now AI (Morgan & Morgan, 14 July 2026; Law.com / Daily Business Review).

Kirkland’s half-billion proprietary build

American Lawyer reported that Kirkland & Ellis is developing a proprietary AI platform and is prepared to spend about $500 million on its AI build over the next three to four years, with input from roughly 250 Kirkland lawyers including about 100 partners. The same reporting notes the firm’s earlier SideTrack tool for investment-fund attorneys (Law.com / American Lawyer, 28 May 2026).

Goodwin’s Regina OS

Goodwin told Law360 Pulse it had developed Regina OS, a proprietary internal AI platform named after former chair Regina Pisa, intended to build and scale practice-specific tools (starting with venture financing) under consistent governance. Chief Digital and Technology Officer Eric Tan described it as a foundation for secure data, workflows, models, and user experiences rather than a single application — built through “Goodwin Builders,” pairing attorneys and technologists, and drawing on Microsoft, Anthropic, and OpenAI partnerships (Goodwin announcement summarizing Law360 Pulse, Aug 2026).

BakerHostetler’s Practice Intelligence Center

Legaltech News reported that BakerHostetler launched its Practice Intelligence Center, described as a firm-built platform combining enterprise data and institutional knowledge with AI-powered capabilities. Launch features include Matter Research+, developed in partnership with Clio and designed to fold the firm’s proprietary matter intelligence into AI research workflows (Law.com / Legaltech News, 28 Aug 2026).

HSF Kramer’s Sovereign System of Intelligence

Legal IT Insider reported that Herbert Smith Freehills Kramer partnered with Acora and Microsoft to build a proprietary Sovereign System of Intelligence on Azure / Microsoft Foundry, governed through Purview, aimed at digitizing firm expertise and interrogating unstructured matter data through an applied ontology (Legal IT Insider, 26 June 2026).

Freshfields Lab: proprietary platforms on frontier models

Freshfields’ public announcements describe proprietary platform solutions developed in the Freshfields Lab that integrate Anthropic’s Claude and Google’s Gemini across firm workflows — Dynamic Due Diligence (D3), a case-management platform, multi-jurisdictional insights, and a general AI platform — while emphasizing co-development rather than building foundation models in-house (Freshfields / Anthropic, 23 Apr 2026; Freshfields / Google Cloud, Apr 2026).

Clifford Chance’s Microsoft-native knowledge bank

Epiq’s case study and Legal IT Insider’s Orange Rag coverage describe Clifford Chance building a custom, Microsoft-native knowledge platform with Epiq and Microsoft — SharePoint, Power Platform, Graph API, Azure AI Studio — delivered in about six months with permissioned search over a curated base of hundreds of thousands of documents (Epiq case study; Legal IT Insider, 21 Aug 2026).

Cleary Gottlieb acquires Springbok to build in-house

Reuters and Cleary’s own announcement report that Cleary Gottlieb acquired Springbok AI — including the SpringLaw platform and a team of data scientists — specifically so the firm could create custom AI-powered solutions through a new AI Acceleration team (Reuters, 17 Mar 2025; Cleary Gottlieb; Legal IT Insider).

Addleshaw Goddard’s AGPT

Addleshaw Goddard announced AGPT as an internally developed, Azure-hosted generative AI platform for secure querying, document review, and translation, developed by its Innovation & Legal Technology and IT teams (Addleshaw Goddard, 2023). Later firm statements have described AGPT remaining in active use alongside other tools, with reported high internal adoption (Addleshaw Goddard, 2025).

Mishcon de Reya’s deReyAI

Mishcon associates and firm materials describe deReyAI as the firm’s private, in-house generative AI suite — private chat, document interrogation, prompt libraries — built for secure early adoption (The Law Society interview; Mishcon; Mishcon innovative client service).

What the pattern suggests

Across plaintiff giants and Magic Circle / Am Law platforms, the recurring move is the same: treat models as commodities and treat firm data, workflows, and evaluation as the product. The sources above disagree on budgets and architecture, but they agree on ownership — these are systems the firms control, brand, and iterate, not merely seats on someone else’s roadmap.

Sources (selected)

This article is a secondary digest of publicly reported material. It is not legal advice.