Casetext (Thomson Reuters) Deep Dive — From Legal AI Pioneer to $650M Acquisition

Opening

March 2023: GPT-4 launches. That same month, Casetext releases CoCounsel — the legal industry's first GPT-4-powered AI assistant. Four months later, Thomson Reuters acquires Casetext for $650 million in cash. By February 2026, CoCounsel's user base has surpassed one million, spanning 107 countries.

This is a textbook case study: a startup built the right product at the right time, then was acquired at a premium by an industry giant, with its brand and technology folded into a much larger ecosystem. I've tracked CoCounsel's product evolution for over two years. This article breaks down the full story.

The Problem They Solve

Legal research is one of a lawyer's most time-consuming tasks. A litigator conducting case research might spend 8 to 15 hours searching for precedents on Westlaw or LexisNexis, reading through cases, and extracting key arguments. Contract review, case summaries, deposition preparation — all of it demands extensive reading and information extraction.

Casetext's original positioning was "a better legal research tool," using AI to improve upon traditional keyword search. The launch of CoCounsel elevated that positioning to "AI legal assistant" — not just helping you find information, but helping you complete entire legal tasks.

Target customer: everyone from solo practitioners to AmLaw 100 firms — a much broader range than Harvey. CoCounsel has no minimum seat requirement, and at $225 per user per month, even small and mid-sized firms can afford it.

Product Matrix

Core Products

CoCounsel Core: A document-processing AI with core capabilities including:

• Document search and review: upload large volumes of files, and the AI extracts key information based on the lawyer's questions
• Case summary generation: automatically summarizes lengthy documents and deposition transcripts
• Case timeline: extracts events from documents and automatically builds a chronological timeline
• Contract analysis: identifies key clauses, risk points, and related provisions within contracts
• Complaint drafting: drafts legal pleadings based on case facts and legal grounds

CoCounsel Legal (launched August 2025): A next-generation platform combining agentic workflows with deep research capabilities. The key upgrade is "guided workflows" — instead of asking lawyers to write prompts, it walks them through structured processes for complex tasks like drafting complaints, developing employee policies, or conducting cross-jurisdictional legal research.

Westlaw Integration: CoCounsel is deeply embedded in the Westlaw ecosystem. Every AI-generated answer is grounded in Westlaw's authoritative legal database, ensuring that cited cases are real and currently in force. This addresses the biggest trust issue in legal AI — hallucination.

Technical Differentiation

After the Thomson Reuters acquisition, Casetext's technology direction shifted significantly:

First, moving from OpenAI dependency to proprietary models. Thomson Reuters is developing its own large language models specifically for legal, tax, and compliance use cases, with the goal of reducing reliance on external model providers.

Second, the data moat. Westlaw's legal database is one of the most comprehensive in the world, covering case law, statutes, legal commentary, and more. Every CoCounsel answer is anchored to this data source — an advantage that independent AI companies have a very hard time replicating.

Third, validation at scale. Usage data from one million users gives Thomson Reuters an extensive foundation for verifying the accuracy and reliability of legal AI.

Business Model

Pricing Strategy

Plan	Price	Target Customer
CoCounsel Core	$225/user/month	Solo practitioners to mid-sized firms
Westlaw Precision (includes CoCounsel)	Bundled with Westlaw subscription	Firms already on Westlaw
CoCounsel Legal	Custom pricing	Large firms / corporate legal

CoCounsel's pricing strategy stands in sharp contrast to Harvey's: no minimum seat count — a single lawyer can subscribe. This dramatically lowers the adoption barrier.

Revenue Model

CoCounsel's revenue is folded into Thomson Reuters' legal business segment. Thomson Reuters invests over $200 million per year in AI productization and has stated that approximately $11 billion in capital is earmarked for continued investment and acquisitions through 2028.

From the acquirer's perspective, paying $650 million for Casetext and seeing CoCounsel reach one million users three years later — the ROI on that deal has been impressive.

Funding and Valuation

Casetext's pre-acquisition funding history:

Date	Event	Amount
2013	Founded	—
2013–2022	Multiple rounds	~$65M total
Mar 2023	CoCounsel launch	—
Jun 2023	Thomson Reuters acquisition	$650M cash

$65 million in funding turned into a $650 million exit — a 10x return. By any measure, a successful outcome for investors.

Customers and Market

Marquee Clients

CoCounsel covers 107 countries and over one million users. Since it's integrated into the Westlaw ecosystem, CoCounsel's clients are effectively Westlaw's clients — the world's largest legal information platform.

Notable users include numerous AmLaw 100 firms and the legal advisory divisions of Big Four accounting firms. Before the acquisition, Casetext already served over 10,000 law firms.

Market Size

The legal research and analytics tools market is approximately $15–20 billion. Thomson Reuters and RELX (parent company of LexisNexis) together hold 60 to 70 percent of this market. CoCounsel's AI capabilities give Thomson Reuters a new growth vector and pricing power within this existing market.

Competitive Landscape

Dimension	CoCounsel (TR)	Harvey	Lexis+ AI	Spellbook
Underlying data	Westlaw (exclusive)	LexisNexis partnership	LexisNexis (exclusive)	No proprietary source
User base	1 million	100,000	Undisclosed	Primarily small firms
Minimum purchase	1 user	25 users	Bundled with Lexis subscription	1 user
Monthly fee	$225/user	$100–500/user	Bundled pricing	From $99/user
Product depth	Medium-high	High	Medium	Medium
Proprietary model	In development	Fine-tuned GPT-4	Partially proprietary	Relies on GPT-4

CoCounsel's core competitive advantage isn't the AI technology itself — Harvey's product may be more advanced. Its edge lies in distribution and the data moat. Westlaw is a standard tool at most law firms, and CoCounsel is embedded within it, making adoption friction close to zero.

What I've Actually Seen

The good: CoCounsel's accuracy is among the best in legal AI, for a simple reason — every answer is anchored in Westlaw's data. In the feedback I've seen, lawyers overwhelmingly approve of the "cites real case law" aspect. One million users' worth of data doesn't lie; this product genuinely solves real needs.

The complicated: Casetext's independent brand was officially retired in April 2025 and fully absorbed into the Thomson Reuters ecosystem. This means that if you're not a Westlaw user, CoCounsel has become less accessible. Additionally, Thomson Reuters' pricing structure is opaque — CoCounsel is bundled across various Westlaw subscription tiers, making the actual cost hard to calculate.

The reality: Thomson Reuters is a company with $7 billion in annual revenue. No matter how strong CoCounsel's AI capabilities are, its pace of development will be shaped by the product cycles and internal priorities of a large enterprise. Several lawyers I've spoken with noted that CoCounsel's iteration speed is noticeably slower than Harvey's. This is the classic startup-vs.-incumbent tension.

My Verdict

• Yes, if: You're already on Westlaw — CoCounsel is a natural extension with the lowest adoption cost. You're a small or mid-sized firm that needs a legal AI tool without a high entry barrier. You're a lawyer who demands the highest reliability from data sources.

• Skip if: You need cutting-edge AI capabilities and the fastest product iteration — Harvey is stronger on both fronts. You're not on Westlaw and don't plan to switch — CoCounsel's value is tightly bound to the Westlaw ecosystem. Your primary need is contract drafting — Spellbook or Ironclad may be more focused.

Casetext's story is a template for startup exits in the AI era: seize the technology window quickly, nail product-market fit, then accept a strategic acquisition while the valuation makes sense. CoCounsel will continue to grow within the Thomson Reuters ecosystem, but its role has shifted from independent product to AI engine powering a larger platform.

Discussion

Do you think Casetext should have stayed independent or was the acquisition the optimal move? Does $650 million look like a bargain in hindsight, or was it fair at the time? If Harvey received an acquisition offer, do you think they'd take it?