Portfolio Approach to Legal AI
$100M/year Harvey for the rest of the world
This shows legal AI buying is turning into portfolio management, not winner take all software selection. Large firms are keeping both Harvey and Legora because each is better at different jobs, then reallocating a small pool of expensive seats to the lawyers on live matters. Harvey pulls demand through US brand recognition and client requests, while Legora is stronger in international workflows, collaboration, and reusable knowledge work.
-
The seat hot swap behavior is a pricing tell. Firms are not buying hundreds of always on licenses. They are starting with 5 to 20 seats, tracking usage closely, and moving licenses from one lawyer or project to the next so high priced tools never sit idle.
-
The product split is concrete. Harvey is stronger for legal reasoning, drafting quality, and US pull from clients. Legora is stronger for parallel agent workflows, contract collaboration, broader team usability, international jurisdictions, and larger document knowledge vaults.
-
This leaves room for a third layer of competition. Clio can bundle legal research through vLex that Harvey and Legora still need to access separately, while specialists like Spellbook, Ironclad, Icertis, and DeepJudge own narrower workflows or the underlying document and search layer.
The market is heading toward mixed legal AI stacks inside the same firm. General tools like Harvey and Legora will stay in the budget, but mostly at practice group level. The biggest upside goes to products that either own the underlying research and document systems, or become the best tool for one daily workflow that lawyers use constantly.