← The Intake
The Intake · Weekly briefing

This week in legal AI

Google Cloud entered legal AI this week, launching Gemini Enterprise for Legal with Cleary Gottlieb, Freshfields, Weil and Williams & Connolly as first customers, while Thomson Reuters shipped its own proprietary model. In Los Angeles, a firm defending State Farm admitted seven fabricated case citations from unsupervised AI use, and the New York City Bar told lawyers AI can assist but never substitute for judgment.

Week of August 22 – 28 2026
Category Market intelligence
Reading time 8 minutes
01 · The week at a glance

Two platforms launched into legal AI, and one firm's use went unsupervised

August 24 · Product
Built on an adapted version of Alibaba's open-source Qwen and trained on decades of Westlaw, Practical Law, Checkpoint and Reuters content, at a reported $40M investment. It initially powers only CoCounsel Legal's tabular-analysis feature.
August 24 · Product
70 tools let any MCP-compatible AI reach a legal team's matter containers (deadlines, documents, meeting transcripts) direct from M365, continuing the interoperability push this briefing has tracked since early August.
August 25 · Market
Cleary Gottlieb, Freshfields, Weil, and Williams & Connolly are named launch customers. The product is available in preview, packaged alongside a parallel financial-services edition.
August 25 · Enforcement
Musick, Peeler & Garrett's lead trial counsel apologizes to a Los Angeles court after opposing counsel's review found seven fabricated cases, nine fabricated quotes, and thirteen fabricated holdings in AI-assisted filings.
August 2026 · Governance
The Emerging Companies & Venture Capital Committee concludes AI tools may assist legal work but cannot substitute for professional judgment.
02 · Market

Google entered legal AI with four elite law firms already signed

Google Cloud spent years selling infrastructure to law firms from a distance. On August 25 it launched Gemini Enterprise for Legal, a purpose-built version of its enterprise AI platform, with Cleary Gottlieb, Freshfields, Weil, and Williams & Connolly named as its first customers. A hyperscaler entering legal AI with four of the most prestigious firms in the world already attached to the launch is a different kind of announcement than a specialist startup's next funding round.

The product packages specialized skills for contract review, due diligence, citation verification, regulatory monitoring, and data-subject access requests, on top of the general Gemini Enterprise platform. Connectors are built to automatically inherit a firm's existing ethical walls from its document and matter management systems, and research outputs are grounded in primary legal authority rather than model training data alone. It launched in preview, alongside a parallel financial-services edition of Gemini Enterprise the same day, with Google describing healthcare and life sciences versions as on the horizon. Notably, the product's Model Context Protocol integrations reach directly into the tools a legal team is likely already running.

Cleary Gottlieb
Launch firm
Freshfields
Launch firm
Weil
Launch firm
Williams & Connolly
Launch firm
Reaches into, via MCP
Harvey Legora iManage Thomson Reuters RelativityOne NetDocuments Docusign Everlaw CourtListener Solve Intelligence Courtroom5
The structural question for enterprise legal teams

Google's connectors are built to reach into Harvey, Legora, iManage and a dozen other platforms a legal team may already run: the same list of tools every other vendor is also racing to plug into. Once every major platform claims compatibility with everything else, the question stops being which platform to buy and becomes something none of them answer: who decides what task goes where, and who signs off before it reaches a client?

03 · Product

Thomson Reuters built its own model to stop depending on one AI vendor

Thomson Reuters spent a reported $40 million training Thomson, a proprietary large language model built on an adapted version of Alibaba's open-source Qwen, rather than routing every CoCounsel Legal query through a frontier lab's API. The company that owns Westlaw and Practical Law just demonstrated that even the largest legal data holder in the world doesn't think one AI vendor should run its whole platform.

Thomson sits on top of an intermediate model Thomson Reuters calls Snowdon, reworked with Imperial College London to be, in the company's framing, ethically and politically de-biased. It's trained on decades of proprietary content from Westlaw, Practical Law, Checkpoint and Reuters, and Thomson Reuters says it shows a meaningful uplift over its open-source base in following complex, multi-part instructions and in reasoning over dense, domain-specific text. For now, Thomson powers only the tabular-analysis feature inside CoCounsel Legal. Everywhere else, Thomson Reuters says it will keep deploying frontier models from OpenAI and Anthropic where they hold a clear advantage. That is a deliberate multi-model routing strategy, not a wholesale replacement of one vendor with another.

Tabular analysis & structured extraction
ThomsonProprietary, TR-owned
Multi-step legal research & drafting
Frontier modelsOpenAI, Anthropic
Everything else in CoCounsel Legal, for now
Frontier modelsPending Thomson expansion
The Flank read

Thomson Reuters isn't trying to replace OpenAI and Anthropic. It's deciding, task by task, which model is worth the cost for that specific class of work, and building the routing logic to keep making that call automatically. That's the same discipline enterprise legal teams need one layer up, at the level of the work itself: a procurement NDA and a novel cross-border restructuring memo aren't the same class of work, and routing both to the same expensive resource, whether that resource is outside counsel or a frontier model API by default, is the waste this briefing keeps coming back to. Inexpensive work done by expensive resources isn't a model-selection problem. It's a routing problem, and it exists whether or not a company can afford to train its own LLM.

04 · Enforcement

A law firm's routine filings for State Farm turned out to be full of fake cases

Musick, Peeler & Garrett was defending State Farm in a Los Angeles fire-loss suit when opposing counsel found the firm's motions in limine cited cases, quotes, and legal holdings that don't exist. Motions in limine are exactly the kind of routine, high-volume litigation drafting billed at a law firm's full rate, and here nobody caught the fabrication before it reached the court.

Plaintiff's attorney Eric Khodadian's review of the filings in Meni-Siliga v. State Farm (a suit stemming from a 2020 house fire in Carson, California, first filed in 2024) turned up seven fabricated cases, nine fabricated quotes, and thirteen fabricated holdings or statements of law. Lead trial counsel Kenneth Katel told the court a co-counsel had used AI to help prepare some of the filings, apologized to the court and the plaintiff, and said he accepted full responsibility as lead counsel. It adds a name, a firm, and a specific list of fabrications to a hallucination tracker this briefing noted in July had already logged more than 1,598 documented cases worldwide.

1
2024
Fire-loss suit filed against State Farm
Filed in Los Angeles Superior Court over a 2020 house fire that made the plaintiff's Carson, California home uninhabitable.
2
This year
Musick Peeler files motions in limine for State Farm
Prepared with AI assistance from co-counsel, per lead trial counsel's later account to the court.
3
Week of August 18
Opposing counsel's review turns up the fabrications
Plaintiff's attorney Eric Khodadian identifies seven fabricated cases, nine fabricated quotes, and thirteen fabricated holdings across the filings.
4
August 25
Lead counsel admits the fabrications and apologizes
Kenneth Katel accepts responsibility to the court; a sanctions request from opposing counsel is pending.
The Flank read

A carrier the size of State Farm pays a firm like Musick, Peeler & Garrett full litigation rates to produce filings like these, and the drafting still went through an AI tool with no independent check before it reached the court. That's inexpensive work, done by expensive resources, with no supervision layer in between; the fabrication wasn't caught by the firm's own process, it was caught by the other side. Flank's answer to exactly this failure mode is structural, not aspirational: agents that know a team's own templates and escalation rules, with a human reviewing the output before anything leaves the system, not a policy that assumes a busy litigator will catch it themselves.

05 · Governance

A bar committee said AI can help lawyers but can't replace their judgment

The Emerging Companies & Venture Capital Committee of the New York City Bar Association published a policy paper this month concluding that AI tools may assist legal work but cannot substitute for a lawyer's professional judgment. The paper names where AI is already embedded (drafting, document review, research, and workflow management) and says the profession still lacks a uniform framework for using it responsibly there.

Written for lawyers and business executives in the emerging-companies and venture-capital space, the paper sets out how existing rules of professional conduct apply to AI use, offers a framework for deciding which legal work is suitable for AI assistance, and recommends steps firms and institutions should take now. It stops short, notably, of prescribing what that framework looks like in operational terms: who checks the output, on what schedule, before it reaches a client.

Where the paper says AI is already embedded
  • Drafting
  • Document review
  • Legal research
  • Workflow management
What the committee says can't be delegated
  • Professional legal judgment, across all of the above, without exception
The structural question for enterprise legal teams

The committee is right that AI can't substitute for judgment, but it doesn't say where in a workflow that judgment gets exercised, or by whom, once AI is producing the first draft of everything from an NDA to a due-diligence memo. A principle without an operational checkpoint is something a team can point to after something goes wrong, not something that prevents it. Where does your own review gate sit, and is it the same gate for a routine NDA as for a novel structuring question?

06 · So what

This week's stories all describe someone else deciding how AI gets checked

Google and Thomson Reuters both shipped legal AI infrastructure this week without saying who reviews the output before a client sees it. A law firm's own filings show what happens when nobody does. A bar committee's answer to the same question was a principle, not a mechanism. Every story this week is a variation on the same missing layer.

A hyperscaler entered legal AI with four elite firms attached
Google Cloud's launch is a distribution story, not a supervision story: its MCP connectors reach into every tool a legal team already runs, without changing who signs off on the output.
Even Thomson Reuters won't bet the platform on one model
Thomson's task-by-task routing between a proprietary model and frontier labs is the same discipline legal teams need applied to their own work, not just their AI vendor's stack.
A law firm's AI use went unsupervised until opposing counsel caught it
Musick, Peeler & Garrett's motions in limine show what happens when routine, high-volume drafting has no review gate attached.
A bar committee named the problem, not the fix
The NYC Bar's "AI can't substitute for judgment" is correct and incomplete without saying where that judgment gets applied in a live workflow.
The Flank view

Take the week's threads together: two platforms added capability, one law firm proved unsupervised capability is a liability, and one bar committee said the quiet part, that judgment can't be delegated, without saying how a firm operationalizes that day to day. None of that is routing infrastructure a legal team owns for itself.

Outsource legal work to supervised agents and the review gate stops being a principle a bar committee states and a firm hopes gets followed. It becomes an actual mechanism: agents that know a team's own templates, terms and escalation rules, with a human checking the output before anything leaves the building. Inexpensive work stops being done by expensive, unsupervised resources. That's the gap every story this week left open.

Subscribe

The Intake

Weekly briefings on what's actually changing in legal AI: the market shifts, regulatory moves, and structural questions that matter for enterprise legal teams. Written by the Flank team.

Subscribe on Substack
Flank

Outsource legal work to supervised agents

Enterprise legal teams use Flank to handle high-volume contracting end-to-end: NDAs, MSA redlines, procurement, triage. Agents that know your templates, terms, and escalation rules. Lawyers review finished work.

Learn more at flank.ai