Legal AI Fragmenting by Practice Area
$100M/year Harvey for the rest of the world
This split means legal AI is turning into a set of narrow work systems, not one firmwide winner. Transactional work rewards speed, pattern matching, and first draft generation, while litigation still depends more on judgment, style, and case specific reasoning. That is why large firms increasingly run Harvey and Legora alongside specialist tools, then move limited seats between teams instead of standardizing on one stack.
-
Inside large firms, both Harvey and Legora usually land as small practice group deployments, not hundreds of seats. Innovation teams monitor usage closely and hot swap licenses across matters, which keeps buyers flexible and makes room for more than one vendor inside the same firm.
-
The deepest product split is between transactional and litigation work. Contract tools like Spellbook fit where lawyers want standardized language, redlines in Microsoft Word, and high volume review. Litigation beyond eDiscovery is harder to automate because briefs and arguments need more originality and matter specific judgment.
-
Practice area specialists are also pulling the market apart from below. Filevine is built around litigation workflows like case management, billing, and plaintiff firm operations, while Clio bundles practice management with legal research through vLex. That leaves Harvey and Legora fighting for the front office layer while incumbents and specialists own narrower workflows.
The next phase is a barbell market. A few broad legal AI platforms will stay in the buying set for general research, drafting, and agent workflows, while more of the durable usage and budget will settle into products built for one motion, like contracts, plaintiff litigation, diligence, or legal research. The winner will be the tool that becomes part of daily work in that exact lane.