Shipping an agent to production means answering questions most tutorials never ask. Where does your agent store its conversation history and uploaded files — in Microsoft-managed infrastructure, or in resources you own and control?

That choice — multi-tenant versus single-tenant — shapes your data residency, compliance posture, and security design from day one.

This full-day session opens with exactly those internals. You will understand how Foundry agents manage state, what single-tenant setup actually requires, and why it matters for enterprise environments.

From there, the session tackles security: how Foundry's built-in guardrails protect against jailbreaking and prompt injection at the model level, and how APIM adds a second, independent defence layer at the API gateway — enforcing content safety, rate limits, and governance before requests ever reach your agent.

The second half of the day shifts to orchestration. You will build multi-agent workflows visually in Foundry, then go code-first with Agent Framework — learning when platform tools are not enough and how to take full programmatic control over agent behaviour, tool integration, and multi-agent coordination. The day closes with Hosted Agents: deploying the full orchestration you built into a production-grade, managed runtime.