video1.0<iframe src="https://www.loom.com/embed/74d52c9f0bbd4df6ad7f502f5e3a3dc0" frameborder="0" width="1692" height="1269" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>12691692Loomhttps://www.loom.com12691692https://cdn.loom.com/sessions/thumbnails/74d52c9f0bbd4df6ad7f502f5e3a3dc0-e51df42b0befd14e.gif431.867This Loom shares four durable frameworks for working with AI effectively in business. Andrew Erickson recommends always inviting AI to the table by running an “AI test” each quarter to see where AI can and cannot help, noting an example where AI is not good at container filling unless you give it explicit math. He emphasizes being the human in the loop and that humans own and can question AI outputs, using restock recommendations as an example where seasonality changes results. He also advises treating AI like a specific kind of person by giving it clear success criteria, and assumes the AI you use is the worst it will ever be since performance improves quickly.