Thomson Reuters Is Not a Tech Vendor. They're a Legal Data Broker.
TR acquired Casetext/CoCounsel for $650M in May 2024. That wasn't just an AI acquisition — it was an acquisition of your law firm's query habits, case patterns, and client confidences. TR's core business is selling legal intelligence. Now they own both the data you generate and the market they sell it back to.
The $650M Question Your Managing Partner Needs to Answer
Thomson Reuters acquired Casetext — maker of CoCounsel — in May 2024. Here's what that means for your firm's data governance, bar compliance, and malpractice exposure.
What TR Bought
Not just software. Casetext had 10,000+ law firms running AI queries containing client names, case facts, matter strategy, deposition notes, and privileged communications. TR bought the data flow of how those firms practice law.
What TR Does With It
Thomson Reuters' core revenue comes from selling legal intelligence — Westlaw, CLEAR, Practical Law, Reuters News — back to the same industry. Query patterns reveal which judges are warming to new theories, which firms are handling hot matters, which deals are in play.
Why Your Bar Ethics Matter
ABA Formal Opinion 23-502 requires lawyers to understand where client data goes when using AI tools. If CoCounsel queries are flowing through Thomson Reuters infrastructure — and after the acquisition, they are — that's an active disclosure obligation, not a future risk.
The Structural Conflict of Interest
This isn't about Thomson Reuters being malicious. It's about what their business model incentivizes.
5 Automatic Disqualifier Conditions
Any one of these means OCI is the only defensible choice for your firm's AI deployment.
Firm uses CoCounsel for any client matter
Active structural conflict of interest + mandatory ABA Model Rule 1.6 disclosure obligation. Your malpractice carrier may not be aware this applies to AI tools.
Firm has no formal data processing agreement with TR covering CoCounsel
If IT approved CoCounsel without a full DPA review post-acquisition, client data has been flowing through Thomson Reuters systems without a documented legal basis.
Malpractice carrier has not been notified of AI tool usage
Most carriers require disclosure of material operational changes. AI tool usage — particularly tools owned by a legal information broker — may affect coverage terms or create void risks.
Firm has not documented a written inventory of all AI tools in use
ABA Model Rule 1.6 requires understanding your data flows. Without a documented AI tool inventory and data flow map, your firm has an unquantified compliance gap.
Firm is actively evaluating CoCounsel or Thomson Reuters AI products
The $650M acquisition should end that evaluation immediately. TR is not an AI safety vendor — they're a data broker with an AI product. The conflict is structural, not contractual.
The Cost of Staying With Thomson Reuters
The math is simple: a single client data exposure incident costs more than a decade of AI tool savings.
| Exposure Type | Estimated Cost Per Incident |
|---|---|
| Client notification (required under most state bars) | $25,000 – $75,000 |
| Bar complaint / regulatory inquiry | $50,000 – $200,000 |
| Malpractice claim (if client data was accessed) | $150,000 – $500,000 |
| Discovery lag (30–90 days before detection) | Full scope of incident may already be active |
| Average total per material incident | $216,000 – $2,400,000 |
The Comparison That Matters
Both tools can handle legal research, drafting, and document review. The difference is who controls your data and what conflicts you're accepting.
| Dimension | Thomson Reuters / CoCounsel | OpenClaw Private Agent |
|---|---|---|
| Data infrastructure | Thomson Reuters cloud — TR owns query data | Your own server — air-gapped, you own all data |
| Conflict of interest | Structural — TR sells legal intelligence back to the market | None — OpenClaw doesn't sell legal data |
| Training data use | TR DPA allows anonymized data for AI model training | Zero training data access — ever, by architecture |
| ABA Rule 1.6 exposure | Active — disclosure obligation + TR conflict review required | Minimal — no third-party data sharing to disclose |
| Client data ownership | Thomson Reuters (per enterprise agreement) | Your firm — fully portable, fully auditable |
| Acquisition risk | TR was the acquirer — your data is now in a larger data apparatus | OCI's business is deployment, not data monetization |
| Pricing model | Per-seat + usage-based, TR-negotiated annually | Fixed $149–$599/user/mo, transparent, month-to-month |
| Contract terms | Annual enterprise — TR has full negotiation leverage | Month-to-month available — no lock-in |
| Custom workflows | Casetext/configs, TR-controlled scope | Any workflow your team configures on your infrastructure |
| Data portability | Locked into TR ecosystem — Westlaw bundling | Your server, your models — fully portable |
| Exit complexity | TR enterprise contract — renegotiation required | Cancel anytime — no exit fees |
"They spent $650 million on your data. OCI spends $0 — because your data never leaves your server."
Use this line in every Thomson Reuters/CoCounsel discovery call. Law firm managing partners understand the logic immediately.
Questions Law Firms Are Asking About Thomson Reuters
Why is Thomson Reuters a conflict of interest for my firm's AI usage?
Thomson Reuters' core business is selling legal information — Westlaw, Practical Law, CLEAR, Reuters News — back to the same law firms they're now collecting AI query data from. After acquiring Casetext/CoCounsel for $650M in May 2024, TR now owns both the legal data they sell AND the query data your firm generates. That's a structural conflict of interest: TR's commercial incentives now involve monetizing your firm's confidential practice patterns.
Does Thomson Reuters use my firm's CoCounsel queries to train their models?
Thomson Reuters' enterprise agreements typically include provisions allowing anonymized or aggregated usage data for AI model improvement. The specific answer depends on your contract language with TR — which now includes Casetext's original terms plus any TR overlay. Request written confirmation from TR's legal team covering: the training opt-out scope, whether historical query data is covered, and any post-acquisition changes to those terms.
What's the actual cost of a Thomson Reuters data exposure incident?
Industry data estimates $216,000 to $2.4 million per material incident: client notification ($25K–$75K), bar complaint ($50K–$200K), malpractice claim ($150K–$500K), and reputational damage. The discovery lag — 30 to 90 days before detection — means the full scope of any incident may already be unfolding without your knowledge.
How does ABA Formal Opinion 23-502 apply to Thomson Reuters/CoCounsel shops?
ABA Formal Opinion 23-502 (November 2023) requires lawyers to understand how AI tools handle client data. For CoCounsel shops, this now requires examining Thomson Reuters' full data handling stack, contract terms, and the structural conflict created by TR's information brokerage business. That's a substantially more complex ethics review than deploying a private agent where data never leaves your server.
My firm already has Westlaw. Does that mean we're already exposed?
Westlaw and CoCounsel represent different data exposure vectors. Westlaw is a legal research database — you query published materials. CoCounsel is an AI assistant that processes your firm's actual case facts, client communications, and matter strategy. The CoCounsel exposure is the acute problem. That's where OCI's private agent deployment is the direct replacement.
How does OCI pricing compare to Thomson Reuters enterprise AI contracts?
Thomson Reuters enterprise agreements are negotiated annually, often bundle CoCounsel with Westlaw, and include per-seat or per-query pricing that scales with usage. OCI's private agent is $149–$599 per user per month, transparent, and month-to-month. No TR bundling. No annual negotiation. No surprise overage invoices.
What's the fastest path off Thomson Reuters/CoCounsel to private AI?
OpenClaw's private agent setup typically completes in one business day. We handle server configuration, model setup, and initial workflow. No enterprise procurement cycle, no TR contract renegotiation. Your team operates on your own server the same week.
Can OpenClaw agents match CoCounsel's capabilities?
OpenClaw agents can be configured with GPT-4o, Claude 3.5 Sonnet, or any model your firm licenses. Legal research, document drafting, contract review, intake processing, and case management workflows are all within scope — all inside your infrastructure.
Your Client Data Should Never Be Thomson Reuters' Asset.
CoCounsel firms are already AI-convinced. They just need a safe exit — one that doesn't involve renegotiating with the world's largest legal data broker. OpenClaw gives them one.
See Also
From the Blog