I'm Ken Priore, a lawyer inside technology companies for more than two decades. I write about what AI does to legal work, and this is where I find out whether I'm right. The essays are at kenpriore.com. The papers and the prototypes are here.
A Trust Primitive for Verifiable Autonomous Legal AI Agent Workflows
When an autonomous agent does legal work, nothing today certifies what it did, under what authority, or whether a human reviewed it. This proposes what that record has to be.
How Legal AI Agents Can Prove They Were Loyal
Fiduciary duty doesn't disappear when an AI agent makes the decision. It just becomes harder to prove. Written for the Information Law Institute's workshop on fiduciary duties and AI.
AI hits legal work unevenly. These maps show where it lands first, what protects a company when production gets cheap, and where the money is going.
AI hits in-house legal work unevenly. The judgment gap dwarfs the process gap.
LiveFive durable moats of the post-production economy: trust, context, distribution, taste, liability.
LiveWhere legal-tech funding is concentrating, and what the concentration says.
LiveThe research companion to the Certificate of Action: the record, not the model, gets audited.
LiveSoftware engineering moved from implementation to orchestration. Medicine did it decades ago. In-house legal is next.
LiveThese map what each stage of legal work runs on, at whatever depth you have time for, and the record a department needs before a lawyer can step back.
The 13-stage operating model. Start here.
LiveEight intelligence types, five data states, thirteen steps. For legal and product teams.
RevisingHow it is captured, evaluated and proven.
DraftingWhere an agent's context comes from, and who is accountable for what is in it.
ResearchingWhat it means for an agent to act inside the authority it was given.
ResearchingIt's easy to say the record matters. It's harder to get that through peer review, argue it in front of law faculty, and ship a prototype that seals what an agent did.
Fast, accurate, cheap, and defensible. The interactive walkthrough of the fourth requirement.
LiveThe research home: the Certificate of Action paper (ICML 2026), the poster, the Fiduciary's Footprint.
LiveThe working prototype. Tamper-evidence across a live agent pipeline, sealed as a Certificate of Action.
DemoLive