It's fun to build things. It's fun to showcase shiny new features, and test out new capabilities. But what happens when you need to audit? Or revert? Or god forbid, track the impact of a change on downstream performance?
With agents being the shiny new toy on the block, we seem to have reverted back to the days where we pushed directly to prod.

We'll be providing a bit of order to this chaos by showcasing a framework to integrate Context as part of your SDLC (Software Delivery Life Cycle), allowing to iterate on changes to your agents in a controlled, auditable and secure manner.

Topics we'll be touching upon:

• What is the distinction between context and semantics?

• How can I ensure that my context is linked to my semantics?

• How can I evaluate the impact of the update of my context on downstream performance of agents?