Podcast: Microsoft Foundry Now Builds, Deploys, and Governs AI Agents End to End

On April 23, 2026, I was invited by Ricardo Josué to join Episodio 11 of “Un show para TI”, a conversation focused on how Microsoft Foundry now builds, deploys, and governs AI agents end to end.

This episode is a technical deep dive for developers and architects who want to understand how Foundry has evolved from a model-hosting service into a full lifecycle platform for AI agents — covering multi-agent workflows, hosted agents, one-click Microsoft 365 deployment, and fleet-wide governance.

Watch the episode

What we covered

Microsoft Foundry has introduced a set of capabilities that change how teams build and operate AI agents at scale. In this episode, I break down the key additions and what they mean for real-world adoption.

Multi-agent workflows

Foundry now supports orchestrating multiple agents that collaborate on complex tasks. Instead of building monolithic agents that try to handle everything, teams can design specialized agents that communicate through structured workflows. This brings modularity, testability, and clearer ownership to agent-based systems.

Hosted agents

Foundry provides fully managed infrastructure for running agents. Developers can deploy agents without managing compute, scaling, or networking. Hosted agents handle the operational complexity so teams can focus on agent logic, prompt design, and evaluation.

One-click M365 deployment

Agents built in Foundry can now be deployed directly into Microsoft 365 with a single step. This means agents can surface inside Teams, Outlook, and other M365 surfaces without custom integration work. For enterprise teams, this is a significant reduction in time to value.

Fleet-wide governance

As organizations move from a handful of agents to dozens or hundreds, governance becomes critical. Foundry now provides centralized controls for managing agent permissions, monitoring agent behavior, enforcing content safety policies, and tracking usage across the entire fleet. This is the layer that makes AI agent adoption sustainable at enterprise scale.

Why this matters for developers

The shift is clear: AI agents are moving from experimental demos to production systems. Foundry is positioning itself as the platform that handles the full journey — from prototyping and evaluation to deployment, monitoring, and governance.

For developers, the key takeaways are:

1. Design for composition — Build smaller, focused agents that can be orchestrated together rather than one agent that does everything
2. Leverage hosted infrastructure — Let the platform handle scaling and operations so you can iterate faster on agent quality
3. Ship where users already are — M365 integration means agents can reach users without requiring them to adopt a new tool
4. Plan for governance early — Permissions, safety policies, and monitoring should be part of the design from day one, not bolted on later

Who this episode is for

This content is for developers, architects, and technical leaders evaluating Microsoft Foundry for AI agent workloads. Whether you are building your first agent or scaling an existing fleet, this episode will give you a clear picture of what the platform offers today and how to take advantage of it.