Casetext CoCounsel
By Thomson Reuters
Casetext CoCounsel is an AI legal assistant, now owned by Thomson Reuters, that helps lawyers with legal research, document review, and drafting by answering questions grounded in a database of case law and a firm's own documents. It was…
Definition
Casetext CoCounsel is an AI legal assistant, now owned by Thomson Reuters, that helps lawyers with legal research, document review, and drafting by answering questions grounded in a database of case law and a firm's own documents. It was one of the first legal AI products built directly on top of a large language model with legal-specific tuning, organizing work around discrete, verifiable tasks.
Overview
CoCounsel grew out of Casetext, a legal research company that had spent years building a searchable database of case law before adding a language-model-based assistant on top of it. That history gave the product an advantage general-purpose chat tools lacked: a large, curated legal corpus to ground its answers in, rather than relying solely on what a general model happened to absorb during training, which for legal citations can be outdated or simply wrong. The assistant is organized around discrete legal tasks rather than open-ended chat, such as reviewing a set of documents for a specific issue, summarizing a deposition transcript, or drafting a first version of a discovery response. This task-oriented structure is a deliberate design choice common across legal AI products, since a narrowly scoped task is far easier to validate for correctness than a freeform conversation with no fixed boundaries. Thomson Reuters' acquisition of Casetext folded CoCounsel into a much larger legal information business that already served courts, law firms, and corporate legal departments with research and workflow tools, giving the product distribution and integration advantages a standalone startup would have taken considerably longer to build on its own. As with other legal AI tools, CoCounsel's outputs are expected to be reviewed by a licensed attorney before being relied upon, and the product emphasizes showing its work, citing the source documents or case law behind an answer so a lawyer can verify it rather than accept it on faith. In practice this makes it most useful for researching case law relevant to a specific question, reviewing contracts or discovery documents for a targeted issue, summarizing lengthy transcripts, and drafting first versions of legal correspondence and filings. CoCounsel competes directly with other AI-native legal assistants as well as AI features being added to competing legal research platforms, where the underlying case-law database and the quality of its citations serve as the main point of differentiation, since two products can use similar underlying models but differ sharply in what evidence backs their answers. For a firm weighing CoCounsel against a competing legal AI product, the decision often comes down to how much value that firm places on Thomson Reuters' existing case-law coverage and distribution relationships versus a newer entrant's potentially different feature set or pricing model. It is also worth noting that being folded into a larger incumbent changes a product's incentives somewhat, since a division of a large company may prioritize integration with the parent's existing tools over the faster, narrower iteration cycle a standalone startup might otherwise pursue.
Key Features
- Grounds answers in a curated legal case-law database, not just general training data
- Organizes work around discrete legal tasks rather than open-ended chat
- Reviews large document sets for specific legal issues or fact patterns
- Drafts legal documents such as discovery responses and memos
- Cites source documents and case law behind its answers for verification
- Backed by Thomson Reuters' broader legal research and workflow business