On this Page
Casetext is an online legal research service. It was founded by Jake Heller, Laura Safdie, and Pablo Arredondo in 2013 and went through Y Combinator the same year. Since its founding, Casetext has raised over $39 million, including the latest round, $8.2 million in February 2020.
The first iteration of Casetext was sort of like Genius for caselaw instead of song lyrics. The user interface was clean and great for reading cases, and, like Genius, users could add annotations or longer explanations.
The hypothesis was that crowdsourced annotation could take the place of the labor-intensive, top-down citators of Westlaw and LexisNexis. The Genius-for-caselaw model did not work out. I think there were two reasons:
- Casetext didn’t get enough annotations.
- Not enough people were willing to pay for a premium subscription.
In any case, Casetext eventually shut down its annotation and blogging services and focused on building AI-powered tools. (However, it kept the information-rich, user-friendly interface.)
Embracing AI: CARA & Compose
In 2016, Casetext launched CARA, an AI-powered legal research assistant. CARA works on any uploaded legal document, like a pleading or brief. CARA analyzes the cases cited, finds other authorities on point, and looks for cases that were missed—or left out. In 2017, the American Association of Law Libraries gave CARA its New Product Award for “products that improve access to legal information, the legal research process, or procedures for the technical processing of library materials.”
Since CARA launched, many of Casetext’s competitors have launched similar tools, including EVA from ROSS Intelligence, Clerk from Judicata (now part of Fastcase), Vincent from vLex, Quick Check from Westlaw, and Brief Analyzer from Bloomberg Law. In 2019, on Above the Law, Bob Ambrogi declared that brief analysis “has officially gone mainstream.”
Also in 2019, Casetext launched Compose, an AI-powered brief drafting tool. It’s sort of like supercharged document assembly. Lawyers select a document type, drag and drop arguments and authorities, and Compose assembles the brief. Compose even suggests citations tailored to the facts written by the lawyer. In a fraction of the time it takes to draft a brief by hand, Compose can create a rough draft for a lawyer to polish.
On the back end, Casetext also uses machine learning to screen judicial opinions for citation authority—whether a case has been upheld, limited, or overruled—so that its human reviewers only have to check passages likely to contain overruling language. (Source.)
Disclaimer: I have an advisor relationship with Casetext that includes a small equity interest.
Published on January 16th, 2021. Last updated on May 26th, 2021, by Sam Glover.