Claude Managed Agents: Design Outcomes and Memory Before You Build

Claude Managed Agents are developing fast. This session covers two new features that determine success: outcome rubrics and memory architecture. No technical knowledge required.

In this session, Charlie Cowan walks through the two core design documents every agent needs before development starts — and shows them live in the Anthropic platform and Kowalah agent hub.

What you'll learn:

• How to write an outcome rubric — defining what 'good' looks like for proposals, support tickets, or any repeatable task your agent handles
• The categories of memory (agent, user, product, and more) and how to map up to eight memory stores to your specific use case
• How to produce a clear brief your technical team can build from — no coding required on your part
• Live tour of memory stores in the Anthropic Managed Agents platform and the Kowalah agent hub

Who should attend:

AI builders, agent managers, and business leaders planning their first one or two agents. This is design and planning work — no technical knowledge needed.

What you'll leave with:

• A method for defining outcome rubrics for any agent use case
• A framework for mapping memory store architecture before you build
• Enough language to brief your technical team with confidence