AI in Context, Ep 01: Delegating to agents

​AI in Context is a weekly live session. Every Thursday at 1 PM ET, for an hour, I show what I'm actually doing with AI that week. Claude, ChatGPT, Codex, the rest. Live, not slides.

​

​Episode 1 is about delegating to agents.

​

​Agent building is glorified SOPs. It's SOPs plus a loop, SOPs plus a trigger. The reason we think our processes are simple is that humans are pretty amazing at abstracting away a lot of stuff. Agents can't. So you have to write down what you would never think to say.

​

​What we'll do:

• ​What to hand off and what stays with you

• ​How to write the handoff so an agent can actually do it

• ​One real task, delegated live, start to finish

• ​How I check the work when it comes back

• ​What I'm running right now: multi-agent setups in the cloud

​You'll leave with the one-page delegation brief I use.

​

​How it runs: about 10 minutes on what's new that week, 35 minutes building live, 15 minutes of your questions. It's recorded, and the replay goes up on lemonbrand.io/live the day after.

​

​Simon Bergeron, Lemonbrand