Thu • Sep 17 • 1:15pm – 2:00pm ET
How to Stop AI Agents from Making Stuff Up
Today's AI models make stuff up far less often than a year ago - and AI agents built on them fail more expensively than ever. No contradiction: the better the model, the more we trust it, the bigger the jobs we hand it, and the subtler the mistakes that slip through. You'll learn to judge any agent by the three questions that decide whether its output can be trusted: What is it allowed to do? What is it allowed to trust? What is it allowed to send? You'll leave with real failure stories, the fixes that caught them, and a checklist to run on your own agents the same afternoon.



