When AI Starts Acting: Securing and Governing Agent Fleets in Microsoft 365

Abstract (WHAT, WHY, HOW, WHO)
AI, Security & M365

WHAT: AI, Security & M365
AI is moving from copilots that draft to agents that execute. This session shows how to design, secure, and govern an enterprise “agent fleet” across Microsoft 365, Security, and AI, using a practical blueprint and a real-world scenario.

WHY: The biggest enterprise risk is not model capability, it’s what happens when an agent’s output turns into action under pressure. Without explicit decision rights, auditability, and a control plane, governance becomes visible only after something goes wrong.

HOW: We will walk through a concrete end-to-end example: an AI agent resolves an IT Ticket. You’ll see how agent identity and least privilege are enforced, how sensitive data access is governed, how tool calls are controlled, and how monitoring and incident response work for agent actions.

WHO: Built for M365 admins, security professionals, architects, and developers who want actionable patterns to move from pilots to production (level 200–300).

Enterprise-Compliant AI: Balancing Innovation and Security with Azure AI Foundry, Purview & Defender

In this session, we explore how enterprises can accelerate AI adoption while staying secure and compliant. We’ll showcase a blueprint for building and deploying AI solutions in alignment with regulatory frameworks like the EU AI Act. Learn how Azure AI Foundry helps innovate responsibly, Purview ensures data governance, and Microsoft Defender safeguards your AI environment.

Join this session to gain actionable insights into creating compliant, secure, and trustworthy AI systems that scale seamlessly in enterprise environments.

Learn how to accelerate AI innovation while staying compliant and secure.
Understand how Azure AI Foundry supports scalable, enterprise-grade AI.
Gain practical strategies for integrating Purview and Defender into your AI workflows.
Discover tools and techniques to prepare for emerging AI regulations like the EU AI Act.

Agenda/Outline
1. Introduction
Quick overview: The challenge of balancing innovation with compliance and security in enterprise AI.
Why AI solutions need to adhere to emerging regulations like the EU AI Act.
2. Azure AI Foundry: Innovation Without Compromise
Overview of Azure AI Foundry: building, deploying, and managing AI models at scale.
Use case: Building production-ready AI copilots with governance in mind.
Integration with Azure AI Studio for development and lifecycle management.
3. Microsoft Purview: Data Governance & AI Compliance
Ensuring compliant and trustworthy data usage with Purview.
Demonstrating data lineage, sensitivity labeling, and policy enforcement.
Purview + AI Foundry: How to trace, govern, and audit AI pipelines.
4. Microsoft Defender: Securing the AI Lifecycle
Overview of Microsoft Defender for Cloud and its role in AI security.
Threat detection for AI systems: Monitoring datasets, models, and APIs.
Proactive risk management and incident response for enterprise AI environments.
5. Putting It Together: A Compliance-Ready AI Blueprint
Step-by-step approach to building AI solutions & governance with Foundry, Purview, and Defender.
Real-world scenarios: EU AI Act readiness and enterprise implementation.
6. Q&A and Discussion
Interactive Q&A to address audience challenges and explore use cases.

Build, Govern, and Scale AI - An Introduction to Azure AI Foundry

Azure AI Foundry is Microsoft’s modern platform for building, deploying, governing, and scaling AI solutions, all in one place. In this introductory session, you’ll learn what Azure AI Foundry is, why it exists, and how organizations are using it to move from experiments to secure, production-grade AI.

We’ll walk through the newest capabilities, including integrated model catalogs, prompt flow orchestration, secure data connections, centralized governance, responsible AI tooling, evaluation frameworks, and integration with Microsoft Fabric and Copilot Studio. You’ll see how Foundry supports both traditional data science workflows and GenAI solutions, while keeping security, compliance, and cost management front and center.

This session is designed for business, IT, and technical leaders who want a clear, practical understanding of the Azure AI platform, without needing deep technical expertise.

Learning Objectives

By the end of this session, participants will be able to:

Explain what Azure AI Foundry is and where it fits in the Microsoft AI ecosystem

Understand the latest features that enable secure, governed AI development

Recognize common enterprise use cases and solution patterns

Describe how GenAI apps are built and monitored within Foundry

Identify the benefits for IT, security, data, and business teams

Session type: keynote-style introduction

Level: 100 (conceptual, non-technical)

Suitable for mixed business + technical audiences

Aligns to Microsoft’s Responsible AI and security principles

Azure AI Foundry Explained: The Modern AI Platform for Enterprises

Azure AI Foundry is more than a model hosting environment, it is Microsoft’s end-to-end platform for building, deploying, governing, and improving AI solutions across the enterprise. This Level 200 session goes beyond the basic introduction and takes you inside how Foundry is actually used to design production AI systems, from GenAI copilots to agent-based applications and RAG knowledge tools.

We’ll explore how teams structure workspaces, manage models, integrate data securely, evaluate performance, and monitor applications in production. You’ll see how Foundry connects with Azure AI Search, Fabric, OpenAI, and the broader Microsoft security stack and how it helps organizations scale safely while maintaining control.

If you’re a practitioner, product owner, architect, or AI lead, this session will help you move from “understanding the concept” to “designing with the platform.”

Learning Objectives

Participants will learn to:

Understand how Azure AI Foundry is structured at workspace, project, and resource levels

Design AI application workflows using Foundry tools and orchestration

Apply evaluation and testing workflows to GenAI applications

Use model lifecycle management to support iterative development

Integrate Foundry with key Azure services for enterprise AI

Audience: Product owners, architects, AI engineers, IT strategy leaders

Level: 200 — assumes basic familiarity with GenAI and Azure

Focus Areas: platform architecture, governance, dev workflows, enterprise enablement

Format: Applied conceptual walkthrough (not hands-on)

Outcome: Attendees gain a structured understanding of how Foundry enables enterprise-grade AI delivery

How to Ensure Microsoft GenAI Solutions are Both Responsible and Compliant to the EU AI Act

Description: As generative AI (GenAI) continues to revolutionize industries, ensuring that these solutions are responsible and compliant with regulatory frameworks like the EU AI Act is paramount. In this session, Emilie Lundblad Director of AI & Automation at Hempel, will explore how to build and deploy GenAI applications that align with both responsible AI principles and the stringent requirements of the EU AI Act. Attendees will gain insight into the operationalization of responsible AI through Microsoft's trusted framework and learn best practices for addressing compliance challenges in high-risk AI scenarios.

The session will cover:
- An overview of generative AI and its transformative role in business.
- Key provisions of the EU AI Act, including high-risk AI classifications and prohibited practices.
- Microsoft's approach to responsible AI, ensuring fairness, transparency, and accountability.
- Real-world examples of compliant GenAI implementations in Microsoft Copilot and Azure AI.
- Best practices and steps for aligning AI solutions with the EU AI Act's requirements.

Also booked for DrivingIT in Denemark in November

Ask Me Anything: AI, Foundry, Copilot, Azure, DevOps

Come with your questions and get them answered by the experts.