Kira Systems

Kira Systems was doing AI contract review before it was cool — its machine-learning extraction models have powered M&A due diligence at large firms since the mid-2010s. Now owned by Litera, Kira remains the archetypal diligence tool: feed it a data room and it pulls out the provisions that matter from a library of more than 1,400 pre-trained fields.

What we like

The extraction accuracy is the product of a decade of supervised training, and it shows. For well-defined diligence questions — change-of-control clauses, assignment restrictions, termination rights — Kira is predictable in a way that prompt-driven generative tools still are not. The custom field training is a quiet superpower: firms encode their own reviewers' expertise into models they own. Because Kira has been a BigLaw staple for years, experienced users are easy to find, and the Word and Excel exports fit straight into standard diligence reports.

Where it falls short

The generative AI wave has left Kira looking narrow. It extracts; it does not draft, negotiate, summarize conversationally, or answer follow-up questions the way Harvey or Luminance can. The interface feels like the enterprise software of its era. And it remains enterprise-only, with pricing and procurement to match — no small firm is buying Kira on a whim.

Verdict

If your firm runs high-volume M&A diligence and values predictable, auditable extraction over conversational flexibility, Kira still earns its keep — especially inside an existing Litera stack. But new buyers starting from zero should evaluate Luminance and Harvey first; the market has moved toward platforms that do more than extract, and Kira's moat is narrowing each year.