CoCounsel

What Is CoCounsel?

CoCounsel is Thomson Reuters' flagship AI legal assistant. Its origin story matters: Thomson Reuters acquired Casetext — the maker of CARA AI, the brief-analysis tool beloved by litigators — in 2023 for $650 million. CoCounsel is the product of merging Casetext's AI capabilities with Thomson Reuters' Westlaw legal content ecosystem and enterprise infrastructure.

The acquisition gave Thomson Reuters something no general-purpose AI company has: a combination of frontier AI capabilities with one of the two largest verified legal content databases in the world. Westlaw's content includes not just case law, but attorney-edited headnotes, secondary sources, treatises, law review articles, and the KeyCite citation validation system. When CoCounsel generates a research answer, it draws on this content — not on scraped web text or parametric model memory.

CoCounsel is available in two forms: CoCounsel Core (the AI assistant features available as part of certain Westlaw subscriptions) and CoCounsel (the fuller product with expanded capabilities). The features available vary by subscription tier, which creates some confusion when comparing what different firms have access to. This review covers the fuller CoCounsel product.

Who Should Use CoCounsel?

CoCounsel's ideal customer is a law firm or legal department that: (1) already has a Westlaw enterprise subscription, (2) primarily does litigation or research-heavy work, and (3) wants to add AI capabilities without a separate vendor relationship. The simpler procurement story — add CoCounsel to your existing Westlaw contract — reduces the friction of enterprise legal AI adoption significantly.

Litigators are particularly well served. CoCounsel inherits Casetext's CARA AI capability — uploading a brief and receiving a list of relevant cases the brief may have missed — which is a genuinely powerful research workflow that has no equivalent in Harvey or Lexis AI. For research-heavy litigation practices, this feature alone justifies evaluation.

CoCounsel is a less natural fit for firms primarily on LexisNexis (where Lexis AI is the parallel offering), for heavily transactional practices without significant research needs, or for firms looking for advanced contract review or due diligence capabilities (where Harvey or Kira Systems are stronger).

Core Capabilities, Examined

Legal Research with Westlaw Integration

This is CoCounsel's strongest capability. A research query in CoCounsel returns an answer grounded in Westlaw's content library, with citations that link directly back to the source material in Westlaw. Because the answer is retrieved — not generated from model memory — the citation accuracy is substantially higher than ungrounded AI alternatives.

The workflow is conversational: you ask a research question, CoCounsel produces a structured answer with supporting cases, you follow up with refinements, and the conversation builds a research thread. Attorneys can export research to a draft memo or work product. The Westlaw integration means KeyCite validation is applied to cited cases — you can see at a glance whether a cited case has subsequent negative history.

In structured testing against known legal research queries, CoCounsel produced significantly fewer hallucinated citations than Claude or ChatGPT used for the same questions. It does produce errors on very recent cases not yet indexed in Westlaw, and on narrow jurisdiction-specific questions where Westlaw coverage is thinner. Independent verification remains necessary — but the baseline error rate is meaningfully lower.

Document Review

CoCounsel's document review capability allows attorneys to upload contracts, briefs, or other legal documents and ask questions about them. It can summarize a document, identify key provisions, flag issues against a checklist, and extract specific information across a document set. The accuracy on well-structured legal documents is high; accuracy degrades on unusual formatting, scanned PDFs with OCR errors, or highly negotiated bespoke contracts.

For litigation document review, CoCounsel can help categorize documents, identify responsive materials, and flag privilege concerns. This is less sophisticated than dedicated e-discovery platforms (Relativity AI, for instance) but serves as a useful first-pass review tool for smaller document sets.

Deposition Preparation

Inherited from Casetext, CoCounsel's deposition preparation feature takes a deponent's prior statements, deposition transcripts, and case documents to produce a deposition outline with suggested questions organized by topic. Litigators who have used this feature consistently rate it as one of CoCounsel's most practically useful capabilities — the output quality is strong enough to serve as a genuine starting point rather than a rough draft.

Contract Analysis

CoCounsel's contract analysis features cover standard review tasks: identifying key provisions, comparing contracts against a checklist, flagging missing or non-standard clauses, and extracting obligations and dates. The capability is adequate for routine commercial contracts but less powerful than specialized contract review tools like Spellbook or Kira Systems. Firms with very high contract review volume should evaluate dedicated CLR tools alongside CoCounsel.

Accuracy and Citation Reliability

The most important accuracy advantage CoCounsel has over general-purpose AI is Westlaw integration. When CoCounsel cites a case, that citation is retrieved from Westlaw — it exists. The question is whether the cited case actually says what CoCounsel says it says. In testing, CoCounsel accurately characterizes holdings the majority of the time on well-known precedents, but mischaracterizes nuanced or narrow holdings more frequently than a senior attorney would.

The practical workflow implication: CoCounsel dramatically reduces the risk of citing cases that don't exist (the problem that led to the famous AI sanctions cases), but does not eliminate the risk of mischaracterizing a case's holding. Attorneys still need to read the cases CoCounsel cites, not just trust the characterization.

Pricing and Value

CoCounsel pricing is tied to existing Westlaw subscription structures and is not publicly disclosed. It's available as an add-on to enterprise Westlaw agreements and as a standalone subscription. Firms report pricing in the range of $100–$300 per user per month for the fuller CoCounsel product, depending on firm size and negotiated rates.

For firms already paying Westlaw's enterprise rates, CoCounsel represents incremental cost on top of a significant baseline investment. The value calculation depends heavily on actual usage: a litigation practice where associates use CoCounsel for research and deposition prep daily will see strong ROI; a practice that uses it occasionally for contract questions will see weaker returns.

CoCounsel vs. Its Main Competitors

Against Harvey AI: Harvey has broader capabilities and deeper due diligence features, but requires a separate enterprise relationship and typically costs more. CoCounsel wins on procurement simplicity for Westlaw shops. Against Lexis AI: this is the direct apples-to-apples comparison — both are database-backed AI assistants for large law firms. CoCounsel has an edge for litigation (CARA brief analysis is unique); Lexis AI is comparable on research. Firms typically choose based on their existing database relationship. See the Harvey vs CoCounsel and Lexis vs Westlaw comparisons for full analysis.

Who Should NOT Use CoCounsel

• →Firms primarily on LexisNexis — Lexis AI is the natural parallel, and switching research databases mid-workflow creates friction
• →Solo practitioners and small firms — the pricing model favors enterprise Westlaw subscribers
• →Practices that primarily need contract drafting assistance — Spellbook or Harvey are stronger here
• →Teams needing high-volume e-discovery review — dedicated platforms like Relativity are more appropriate
• →Firms that want to evaluate AI without committing to a Westlaw subscription change