Firms prefer practice-specific drafting tools

This shift means legal AI drafting is breaking into narrow workflow products, not converging into one general copilot. In practice, firms want software that behaves like a specialist associate, it asks the extra questions a real domain lawyer would ask, uses the right precedent set, and produces a first draft that can actually move into review instead of being rewritten from scratch. That is why broad tools often land as research and review layers, while drafting gets carved up by practice area.

• Inside large firms, drafting quality depends heavily on context capture. The strongest tools are built around one legal lane, such as contracts or patents, and guide the user through missing deal terms, fallback positions, and clause logic, instead of waiting for a perfect prompt.
  1 sacra 2 sacra 5 sacra
• Harvey and Legora still matter, but more as broad legal workbenches than universal drafting engines. Firms evaluate them across research, review, and internal knowledge workflows, yet often deploy them in small seat counts by practice group because enterprise wide licensing is too expensive relative to usage.
  1 sacra 3 sacra 4 sacra
• The clearest comparable is Spellbook, which goes deep on contracts inside Microsoft Word rather than trying to cover every legal task. That narrower workflow fit makes it easier to justify on standalone ROI, while Harvey and Legora bundle contract work into a larger platform sale.
  5 sacra 6 sacra

The market is heading toward a split stack. General platforms will own cross practice research, document analysis, and internal knowledge access, while specialist products win the drafting moments where the right follow up question matters more than raw model power. The companies that compound fastest will be the ones that turn legal specialty knowledge into repeatable drafting workflows.