AI legal research and drafting assistant for practicing attorneys
Casetext CoCounsel is an AI-powered legal assistant built on GPT-4 for attorneys.
AI Panel Score
6 AI reviews
Reviewed
Casetext CoCounsel is an AI legal assistant developed by Casetext, a legal technology company, and built on OpenAI's GPT-4 model. It is designed for practicing attorneys and law firms who need to perform research, review documents, and prepare legal materials more efficiently. The tool is trained and configured for legal contexts, distinguishing it from general-purpose AI writing tools.
CoCounsel supports a range of legal tasks including case law research, contract analysis, deposition preparation, legal memo drafting, and document review. Attorneys can input queries or upload documents and receive structured, cited responses that are intended to reflect the standards of professional legal work. The system is designed to surface relevant statutes, case law, and contractual clauses in a format attorneys can act on.
The product is positioned within the legal technology market alongside other AI-assisted research platforms. Casetext had an existing legal research platform before launching CoCounsel, and the assistant integrates with that research infrastructure. In 2023, Casetext was acquired by Thomson Reuters, which has influenced the product's distribution and development roadmap.
CoCounsel is intended primarily for attorneys at law firms, in-house legal departments, and other legal professionals. It is a web-based tool, and access is provided through a subscription model. Pricing has been geared toward professional and firm-level use, reflecting its enterprise focus rather than a consumer audience.
Identifies key issues, risks, and relevant provisions across uploaded contracts; compares terms and flags deviations.
Hands off full legal research questions to an autonomous AI agent that reasons, plans, sources answers, and builds argument foundations.
Generates deposition question lists tailored to a case theory and uploaded evidence.
Summarizes long legal documents and depositions into concise briefs with citations to source passages.
Drafts pleadings, motions, and legal correspondence from natural-language instructions.
Helps prepare trial strategies by analyzing case materials and surfacing argument framings.
Pre-built workflows for drafting privacy policies, employee policies, complaints, and discovery requests/responses.
Builds chronological case timelines from uploaded discovery materials and depositions.
Reviews deposition transcripts to surface inconsistencies, key admissions, and follow-up question candidates.
Searches and reviews uploaded documents; extracts specific provisions across large transactional contract sets.
Connects CoCounsel to Practical Law content for guided workflow templates and best-practice resources.
Grounds research and analysis in Thomson Reuters Westlaw content for verifiable legal citations.
CoCounsel is a professional-grade AI platform for legal, tax, audit, and accounting professionals. No public pricing is listed; prospective customers must contact Thomson Reuters for a quote.
A board-proof legal AI bet from Thomson Reuters, gated only by quote-only pricing.
“CoCounsel is owned by Thomson Reuters and crossed one million users across 107 countries in February 2026. The catch is no public price, so procurement walks in blind.”
One million users in 107 countries, announced February 2026. Thomson Reuters paid $650 million for Casetext back in 2023. No board asks whether this vendor survives three years.
The real call is whether CoCounsel advances your firm or just speeds up research associates already do. Deep Research hands a full question to an autonomous agent that plans, sources, and builds argument foundations grounded in Westlaw content. Harvey competes hard on legal-tuned models, but the differentiator here is the Thomson Reuters corpus the citations verify against, not the chat box.
However, the buying process is the friction. Pricing is contact-sales only, scaled to seat count and content packages, so you cannot model cost before a demo. Reputation risk is near zero with a public-company vendor. Pilot it with one practice group for 90 days, confirm the citation accuracy, then negotiate the firm-wide seat math.
About a quarter of Fortune 1000 companies use it, signaling peers are already in.
A public-company vendor with one million users is an easy choice to defend to the board.
Guided workflows shorten routine drafting, but firm rollout still needs a 90-day pilot.
Deep Research and Westlaw grounding advance legal work rather than just trimming cost.
Owned by publicly traded Thomson Reuters, which paid $650 million for Casetext in 2023.
Law firms who want verifiable AI research backed by a durable vendor.
Solo practitioners who need transparent per-seat pricing upfront.
CoCounsel's moat is Thomson Reuters content, but adopting it means committing to that corpus.
“CoCounsel grounds agentic legal AI in Westlaw and Practical Law, a corpus rivals cannot replicate. The craft ceiling is high, but the value is inseparable from Thomson Reuters' content and quote-only pricing.”
A general counsel choosing a legal-AI substrate for the next three years should weigh what the answers are grounded in, not the chat surface. CoCounsel's bet is content: Thomson Reuters bought Casetext for $650M in 2023 and rewired the assistant onto Westlaw and Practical Law. That grounding is the architecture.
The craft ceiling is real. Deep Research hands a full research question to an autonomous agent that plans and sources every answer against Westlaw citations, while Contract Analysis flags clause deviations across large transactional sets. Against Harvey, which fine-tunes its own models, CoCounsel's edge is verifiable authority rather than model novelty.
The catch is the lock-in. The value lives in Thomson Reuters content, so leaving means leaving the corpus, not just the tool, and pricing is quote-only with no public tiers. With one million users reached in February 2026, this is a durable bet on the incumbent rather than the frontier.
One million users by February 2026 makes it the incumbent platform in legal AI.
Contract Analysis, deposition review, and pleading drafting map directly to how litigators and transactional attorneys work.
Integrates with Westlaw, Practical Law, Checkpoint, and Microsoft 365 across the legal stack.
Adoption ties the firm to Thomson Reuters' content corpus, a durable but deepening dependency.
Westlaw and Practical Law grounding plus Deep Research agentic reasoning is best-in-class craft for legal AI.
Law firms who need AI research grounded in verifiable Westlaw citations.
Solo practitioners who want self-serve evaluation and public pricing.
No published price, contact-sales only, and a quote that scales with seat count and content packages.
“CoCounsel lists nothing — every number comes from a Thomson Reuters sales call. The quote stacks Westlaw and Practical Law content fees on top of seats.”
No sticker on the page. CoCounsel sells through Thomson Reuters as an enterprise quote, priced on firm size, seat count, and content packages. Procurement starts blind. Casetext sold to Thomson Reuters in August 2023 for $650M, so the vendor is durable, not a startup risk.
The catch is the content stack. The AI seat is one line; Westlaw Integration and Practical Law access are separate licenses you likely already carry or must add. A 30-attorney firm bundling Westlaw plus CoCounsel can clear six figures a year before training time. Harvey prices the same opaque way, so blind quotes are a category norm here, not a Thomson Reuters quirk.
ROI is legible. Deep Research and contract review produce auditable billable-hour savings, and one million users as of February 2026 signals a stable roadmap. Negotiate term length and audit content-package overages.
Thomson Reuters is an established vendor with low onboarding risk after the $650M Casetext acquisition.
Enterprise terms leave negotiation room on seats and term length, though no public cancellation policy.
No published tiers — pricing page lists only Contact Sales, so every buyer starts blind.
Deep Research and contract review map directly to auditable billable-hour savings attorneys can measure.
Seat cost stacks on Westlaw and Practical Law content licenses, pushing firm-level spend into six figures.
Law firms who already license Westlaw and want grounded legal AI.
Solo practitioners who need transparent published per-seat pricing.
CoCounsel grounds research in Westlaw and cites every clause, but pricing stays behind a sales call.
“Deep Research and Contract Analysis cut the routine grind out of a litigation associate's week. But there is no public pricing and no free plan, so scoping it solo is off the table.”
An associate measures legal AI by the week a discovery deadline closes in, not the partner demo. CoCounsel's Deep Research hands a research question to an agent that plans, sources, and returns answers grounded in Westlaw and Practical Law. Every holding traces back to a cited authority instead of a confident paragraph you re-verify against good law.
The daily win is Contract Analysis across a document set. Upload a transactional batch, ask it to flag deviations, and it works the whole set at once instead of one PDF per prompt. Lexis+ AI covers similar ground, but CoCounsel's Guided Workflows feel built around how a matter moves from intake to filing.
No public pricing and no free plan, though. Access runs through a Thomson Reuters sales quote keyed to firm size. A solo practitioner can't scope a pilot. With one million users announced February 2026, this is built for firms.
Cited, Westlaw-grounded answers hold up under the verification pressure of a real filing week.
Has docs and a blog; answers cite source passages so practitioners can verify rather than trust.
Batch document handling reduces re-prompting, though no-trial access adds friction before onboarding.
Deep Research agents and twelve task types scale from quick lookups to full trial-strategy prep.
Guided Workflows and document-set Contract Analysis map onto how matters move from intake to filing.
Litigation and transactional attorneys at firms who run research and contract review weekly.
Solo practitioners who need transparent pricing before committing.
CoCounsel earns its keep on cited research, but you cannot see a price without a call
“The Westlaw grounding means answers come back with sources you can actually check. The catch is no public pricing and no free tier, so the first ten minutes are a sales form.”
A million professionals across 107 countries were using CoCounsel by February 2026, so this is no demo-glow gamble — it is a tool people actually keep open. What you feel day three is that the answers come back cited. Ask it something and it points at Westlaw and Practical Law passages, not open-web guesses you then have to babysit.
Deep Research is the part that earns its keep. Hand it a full legal question and the agentic flow plans, sources, and builds an argument foundation while you stay in the loop. Contract Analysis does the same trick across a stack of uploaded agreements, flagging deviations instead of leaving you to scroll. It feels built by people who know the work runs late.
Harvey is the obvious standalone rival. But CoCounsel hides its price behind a Contact Sales form, and since the 2023 Thomson Reuters acquisition it is firmly enterprise-shaped, so you cannot size your real cost without a rep on the line.
Cited answers and structured outputs show a team that sweated the verification details lawyers care about.
Guided Workflows and pre-built flows give attorneys a clear path from first hour to month three.
Web-only enterprise legal tool where mobile is not the core use case, scored neutral.
No free tier and a Contact Sales gate mean the first ten minutes are a quote request, not the product.
Westlaw and Practical Law grounding plus zero-retention API calls and AES-256 encryption signal a solid, trustworthy base.
Attorneys who need cited, verifiable research instead of open-web guesses
Solo practitioners who want to see a price before talking to sales
A Thomson Reuters product with a real corpus behind it, but the version under you keeps changing.
“Casetext was acquired for $650M and the tool now reaches a million users. The catch is contact-only pricing and a product that was relaunched and renamed since launch.”
The vendor-survival question is mostly answered. Casetext got acquired by Thomson Reuters for $650M in 2023, and the company announced one million CoCounsel users in February 2026. A 173-year-old public parent is not the legal-AI graveyard pattern. The startups racing it are.
What I would watch is product churn, not company churn. CoCounsel launched on GPT-4 in March 2023, then was relaunched as CoCounsel Legal in August 2025 with Westlaw-grounded agentic AI. That is a fast rewrite for a tool firms build billing workflows around. The features are concrete — Deep Research hands full questions to an autonomous agent, and Westlaw Integration grounds answers in verifiable citations. Harvey leans on model tuning instead; the corpus here is the moat.
The yellow flag is opacity. Contact-only pricing, no public tiers, no free plan. You cannot benchmark cost before you commit.
Westlaw and Practical Law grounding is a content moat Harvey and Lexis cannot easily copy.
Documents export, but Deep Research workflows and Westlaw-grounded analysis do not port to a rival.
Thomson Reuters backing, a million users, and steady relaunches signal a durable three-year bet.
Claims like one million users and Westlaw grounding are specific and verifiable, not vague aspiration.
A profitable public parent and a real legal corpus match survivor patterns, not the startup graveyard.
Law firms who already pay for Westlaw and want grounded legal AI.
Solo practitioners who need transparent pricing before committing.
Common questions answered by our AI research team
Deep Research is an agentic AI capability that hands off full legal research questions to an autonomous agent. It reasons, plans, and delivers comprehensive results grounded in Westlaw and Practical Law content, sourcing each answer and building argument foundations with human oversight.
Yes. Upload a set of contracts and ask CoCounsel to compare terms, flag deviations, or extract specific provisions. It is built for transactional practices handling contract review across large document sets.
Guided workflows currently include drafting privacy policies, employee policies, complaints, discovery requests and responses, and deposition transcript reviews.
CoCounsel is grounded in Thomson Reuters Westlaw and Practical Law content. Answers cite the underlying sources, which means the legal analysis can be verified against authoritative material rather than open-web data.
CoCounsel is sold as an enterprise legal AI product through Thomson Reuters with no public pricing tiers — pricing depends on firm size, seat count, and content packages. Contact Thomson Reuters sales for a quote.
Casetext built the original CoCounsel on top of OpenAI's GPT-4 in 2023. Thomson Reuters acquired Casetext in August 2023 for $650M. The current product, CoCounsel Legal, was relaunched in August 2025 with Westlaw-grounded agentic AI capabilities.
Practicing attorneys and legal professionals at law firms and corporate legal departments. Use cases extend beyond core legal research to tax law, regulatory compliance, due diligence, and contract review.
Thomson Reuters announced one million CoCounsel users in February 2026, indicating broad adoption across the legal industry.
Company
Thomson ReutersFounded
2008Pricing
Contact for pricingFree Trial
AvailableThomson Reuters is a Toronto-based information and technology company providing Westlaw, Practical Law, Checkpoint, and AI-powered tools for legal, tax, and news professionals.
