The CAFE(S) framework: Improving AI agent effectiveness through better context (APAC)

As AI agents become more capable, context quality is increasingly becoming the limiting factor in their performance. DX hosted an early look at CAFE(S), a new framework for evaluating whether AI agents have the context they need to do their jobs well. The session introduced the five dimensions of CAFE(S) and demonstrated how engineering organizations can use them to understand how context quality influences agent performance, developer productivity, operating costs, and scalability. This session includes Justin Reock (Deputy CTO, DX) and CAFE(S) contributing authors Brian Houck (Distinguished Scientist, DX),and Eirini Kalliamvakou (Research Advisor, GitHub).

The session covers the following:

A practical framework for evaluating the quality of AI agent context.
A shared vocabulary for diagnosing common context failures before they become production problems.
Guidance for individuals, teams and organizations to improve context quality.

Watch the recording

Speakers

Justin Reock

Deputy CTO, DX

Brian Houck

Distinguished Scientist, DX

Eirini Kalliamvakou

Research Advisor, GitHub

Powering the world's top companies