Your Briefing: What's Happening in Legal AI Right Now

From Kirkland's $500M bet on proprietary AI to OpenAI's move into transactional workflows, an infrastructure shift unlocking AI across deal documents, and what surging in-house adoption means for outside counsel. Here's everything you need to know from the last month.

By the Numbers

‍‍• $500M  Kirkland & Ellis's total AI investment commitment, the largest disclosed by any law firm to date

• 87%  of in-house legal teams now use generative AI, nearly double the 44% from a year prior

• 3 hrs  average time lawyers spend reviewing a single contract manually (LegalOn, 2026)

• 32%  of legal organizations cite integration complexity as their biggest AI adoption barrier (iManage, 2026)

Sources: Bloomberg Law  ·  FTI Consulting / Relativity 2026 General Counsel Report  ·  LegalOn 2026 Research  ·  iManage Knowledge Work Benchmark Report 2026

Top Stories

BigLaw

Kirkland bets $500M on proprietary AI. The Palantir deal shows what that looks like in practice

The world's highest-grossing law firm just put the largest publicly disclosed number on its AI ambitions. Kirkland & Ellis announced a $500 million multi-year AI investment program, with more than $100 million earmarked for 2026 alone. The first product out of that program: a partnership with Palantir to build a fund formation engine. It is an AI system designed to handle fund documentation, compliance tracking, obligation monitoring, and investor communications for private equity clients, drawing on the institutional knowledge of Kirkland's 1,000-attorney Investment Funds Group.

The deal is significant for transactional lawyers for a few reasons. First, the sheer scale signals that proprietary AI, built to encode a firm's own knowledge and workflows and unavailable to competitors, is now a serious strategic investment rather than a pilot. Second, the PE focus is deliberate: Kirkland advised on funds raising close to $500 billion last year, with clients including Blackstone, Thoma Bravo, and EQT. Embedding senior partner judgment into an AI system for fund documents and side letters compresses a time-intensive cycle with real implications for how quickly sponsors can close oversubscribed vehicles. Third, Kirkland's architecture is model-agnostic by design, meaning the firm can swap underlying AI models as technology evolves without rebuilding from scratch.

→ Bloomberg Law: Kirkland's $500 Million AI Gambit Requires a Cast of Hundreds

→ PE Insights: Kirkland & Ellis taps Palantir for AI tool aimed at PE fundraising work

Business of Law

‍OpenAI enters legal. Its hire of the Ironclad founder tells you something about their strategy

OpenAI formally launched a dedicated legal vertical in June, following reports in May that it was planning an offering branded 'Codex for Legal.' The person OpenAI chose to lead it is Jason Boehmig, co-founder and former CEO of Ironclad, one of the most widely used contract management platforms in the industry, and a former corporate attorney at Fenwick & West. He is not a model expert. He is a decade-deep expert in how contracts actually move through organizations.

That hire is the whole thesis in miniature. OpenAI already has model capability. What it went out to acquire is workflow knowledge: how legal work routes, how approval processes function, how lawyers and counterparties interact around documents. Anthropic moved first with Claude for Legal (launched in May with 90+ specialized legal agents and 12 plugins), and Microsoft has a Legal Agent in Word. Three of the world's most powerful AI companies now have dedicated legal offerings, all targeting the legal application layer.

For firms evaluating which AI infrastructure to commit to, the more important question is not which tech giant is behind the tool, but whether it was actually built with how legal work gets done in mind. General-purpose models and purpose-built legal platforms are meaningfully different. Analysts are also flagging a vendor lock-in risk worth taking seriously: committing deeply to a single provider limits flexibility as the underlying models continue to evolve rapidly.

→ Artificial Lawyer: OpenAI Plans 'Codex for Legal'

→ Indian Legal Tech: Why OpenAI's Move Into Legal Is a Fight Over Workflow, Not Models

Adoption

In-house AI adoption nearly doubled over the past year. Clients are raising expectations for firms

The FTI Consulting / Relativity 2026 General Counsel Report puts in-house generative AI usage at 87%, up from 44% a year prior. That acceleration is reshaping what corporate legal departments expect from outside counsel. In-house teams that have automated their own contract review, due diligence triage, and drafting workflows are increasingly impatient with firms that have not done the same. The pressure on the adoption side of the market has not relented. If anything, the gap between in-house capabilities and firm capabilities is becoming a more visible issue in panel reviews and instruction decisions.

For transactional practices in particular, the efficiency bar is moving. The firms gaining ground on mandates are those that can show clients AI is making their work faster and more thorough. That case is easiest to make when AI is embedded in the actual drafting and review workflow, not bolted on as a separate step. The firms still treating adoption as optional are finding that clients have stopped waiting.

→ FTI Consulting / Relativity: The General Counsel Report 2026

Tools Worth Watching

iManage opens its MCP server. Deal documents just became a lot more accessible to AI

iManage, the document management platform used by 83% of the Global Top 100 law firms and 79% of the AmLaw 100, launched its MCP Server in May. It makes governed iManage content accessible to any compatible AI tool via a single standardised connection. The practical effect for transactional teams: AI assistants can now access matter history, prior agreements, and institutional precedent held in iManage directly, without bulk data exports, custom integrations, or sacrificing the access controls and ethical walls already in place.

The broader shift MCP represents is worth understanding: legal AI is moving from standalone tools toward connected systems that can draw on the full context of a matter. For transactional teams, that means the AI assistance happening at the document level, where drafting and negotiation actually occur, becomes more powerful when it can reference the broader matter history and precedent sitting in the DMS. The two layers are complementary, and the firms thinking about both are the ones building a coherent AI strategy.

→ LawSites: iManage MCP Server Now Available

→ Artificial Lawyer: MCP: The Standard that Decides Legal AI's Future