I'm the GC building an AI-powered legal function at Worksome and I’m documenting everything.
I asked an AI tool for the top five risks in a contract and got a clean, correct, completely "legal" answer. But the clauses that mattered from an operational perspective were not flagged. Learn more about the categories of risk that most AI contract review skips, and the skill I built to fix it.
I grew up watching Silicon Valley build things because they could, then figure out the consequences later, or not at all. That's the lens I bring to AI now. Not fear, a method. Here's how I actually stress-test what AI tells me, including the skill I built to do it automatically.
Last week Wordsmith changed something and I had to learn a new way of building a solution. Then my husband said ChatGPT had done the same to him. Claude keeps rearranging things too. I stopped asking which tool is best and started asking a different question: what do I own that survives the next update?