Cooley, a law firm that handled over 200 IPOs in 2021 alone, built a tool called GO Public using ChatGPT Work to automate early-stage due diligence in the IPO pipeline. The system surfaces legal issues in draft filings before senior attorneys review them, compressing a process that previously ate billable hours into something a junior associate and an AI can front-load.
The reason this matters is not the time saved. It is where the time goes instead. By delegating initial document review to the model, Cooley's lawyers are repositioning their judgment at the exceptions, the edge cases, and the client-facing decisions that require legal expertise. That is a structural change in how legal work gets allocated, not a productivity tweak.
The full piece is worth reading for the specific workflow details: how GO Public was scoped, what it is actually permitted to flag versus defer, and where Cooley draws the line on model autonomy inside a regulated process. Those constraints reveal more about enterprise AI deployment than the headline outcome does.
[READ ORIGINAL →]