Short takes on legal tech, AI transformation, and product thinking in regulated environments.
Legal AI fails when it's scoped by people who understand the technology but not the legal process it's replacing. Legal reasoning isn't just information retrieval — it's structured judgment under constraint. The products that work are the ones built by teams that understand both sides: what the AI can do, and where a solicitor would push back in court. Most teams skip the second part.
In every transformation programme I've led, the hardest problems weren't technical. They were organisational. Who owns the new workflow? What happens to the team whose job just changed? What does the compliance function think of a model making recommendations? AI transformation that survives contact with reality is transformation that treats change management as a product problem — not a communications afterthought.
Speed and agility are table stakes. In regulated environments, the differentiator is the ability to hold complexity without oversimplifying it. Legal, financial, and compliance-heavy products fail when PMs strip out nuance to make requirements fit a sprint. The skill is translating regulatory constraint into product logic — clearly enough for engineers to build, accurately enough for lawyers or auditors to sign off.