Enterprise AI Development, Year Three. Own Your Agent

For my third year speaking at M365 Community Days NYC on enterprise AI development, I’m bringing the third act. Ownership.

Enterprise teams now need to answer harder questions. Where does the agent run? What can it access? Which skills define its behavior? What data boundaries does it respect? What test proves the behavior still works after the next change?

In this live technical session, live, we will inspect real skill files, run validation checks, execute behavioral tests, and write a new failure-mode eval for a privacy boundary. The architecture uses OpenClaw as the runtime, Tula skills for inspectable domain behavior, Microsoft infrastructure for the enterprise integration layer, and Waza evals for the verification harness.

The goal is simple. Move from impressive Agent demos to owned, inspectable, and verifiable agents that enterprise teams can actually operate.

Own the Runtime - Microsoft Infrastructure for Verifiable Enterprise Agents

Not one model. Not one vendor. One owned runtime.

This session shows how to build an owned agent runtime using Microsoft infrastructure, OpenClaw, and Waza. The reference use case is Tula, an open-source personal health AI built as OpenClaw skills.

We will walk through the architecture for an enterprise-controlled agent running on Azure, using Entra ID and Microsoft Graph for Microsoft 365 integration, OpenClaw for the agent runtime, Tula skills for inspectable domain behavior, and Waza evals for behavioral verification.

The key lesson is not “use one model” or “use one vendor.” The opposite is true, real agents are heterogeneous. Enterprise teams need to own the runtime, scope the permissions, isolate the workspace, and verify the behavior before they trust the agent inside business workflows.

Attendees will leave with a reference pattern for Microsoft-centered agent systems that are inspectable, self-hosted, and testable.

From Vibes to Verifiable with Tula, OpenClaw, and Waza

Tula is an open-source personal health AI built as a set of OpenClaw skills. In this live technical session, we will inspect the actual skill structure, run validation checks, execute behavioral evals, and show how a self-hosted agent can be owned, tested, and improved without disappearing into a vendor black box.

The demo moves through three layers, the OpenClaw runtime, the Tula health skills, and the Waza eval harness. We will look at real SKILL.md files, run Waza checks against the repo, execute behavioral tests for health-record and medical-document workflows, and write one new test from a plausible failure mode: “the agent tries to move protected health information outside the workspace.”

The key idea is simple. Owning your agent does not just mean hosting it yourself. It means owning the runtime, the skills, the data boundary, and the tests that prove the behavior still holds after the next change.

Attendees will leave with a practical pattern for building agent systems that are open, inspectable, and verifiable.

Microsoft’s Agent Stack - Copilot (SaaS), Azure AI Agent Service (PaaS) & MS Agent Framework (IaaS)

Microsoft’s ecosystem now offers a full continuum for intelligent agents, from turnkey Copilot experiences in Microsoft 365 (SaaS), to the new Azure AI Agent Service for plug-and-play agent orchestration (PaaS), down to open-source frameworks like Semantic Kernel and AutoGen that let you craft custom agent patterns at the infrastructure layer (IaaS).

In this fast-paced session, we’ll, demystify the layers - how user-facing Copilots, Azure AI Agent Service, and Microsoft Agent Framework map to SaaS, PaaS, and IaaS design patterns.

Attendees will leave with a concrete framework for selecting the right Microsoft agent technology, battle-tested patterns for rapid deployment, and actionable next steps to create scalable, compliant agent solutions in their own organizations.

Building Next-Gen Products with Microsoft AI Developer Services (Manila Edition)

Tara na. let’s build the future, together.

Join me for an advanced session designed for product managers and architects on the new art of AI applications. Explore the latest AI developer tools from Microsoft, focusing on agentic programming, intelligent orchestration, and conversational AI.

Additionally, gain insights on deploying safe AI products and solutions in the Azure Marketplace, ensuring they meet standards and leverage Microsoft's ecosystem. This session offers essential strategies to lead in the era of advanced AI and transform your product development process.

Demystifying AI Agents for Malaysia’s Next Digital Leap

With Microsoft’s $2.2B cloud investment bringing AI infrastructure to Malaysia and the Madani vision pushing digital adoption, leaders must now move beyond AI hype to real-world deployment. But what exactly are AI agents—and how can they be used safely in mission-critical settings?

In this executive-level session, Paul Swider, a Strategic Microsoft Partner and Thought Leader, demystifies AI agents by breaking down their architecture, strategic value, and role in reshaping services across sectors. Grounded in the Microsoft ecosystem—Copilot, Azure AI, and Semantic Kernel—Paul shares proven models and governance frameworks to help organizations in Malaysia deploy AI agents with confidence.

Key Takeaways:
-Understand what makes AI agents different from chatbots and copilots
-Learn how Microsoft tools (Azure OpenAI, Semantic Kernel) power agent-based systems
-Explore real use cases in healthcare, citizen services, and compliance
-Use the Wheelhouse™ Readiness Framework to assess your organization’s AI maturity
-Align agent deployments with Malaysia’s regulatory and digital goals

Radical AI Governance for Healthcare - How Dubai's is Redefining Global Governance

Dubai welcomed 691,000 medical tourists in 2023, generating AED 1.03 billion in direct healthcare revenue and AED 2.3 billion in GDP contribution. Yet most healthcare providers are still competing on price and facilities, missing the transformative potential of AI to revolutionize the entire patient experience, from discovery to post-treatment follow-up across continents.

In this session, Paul Swider, Chief AI Officer at RealActivity and architect of cross-border healthcare AI systems demonstrates how intelligent orchestration transforms medical tourism from transactional service delivery into relationship-driven, lifetime patient engagement. Drawing from Dubai's unique positioning as a global healthcare crossroads and Microsoft Healthcare AI, you'll discover how AI creates competitive moats that price competition cannot erode.

From Boston to Dubai - AI-Driven Operational Efficiency in Healthcare

Two global healthcare innovation powerhouses are leading the AI transformation in medicine. Boston, with its concentration of world-class teaching hospitals and innovation accelerators like MassChallenge HealthTech, and Dubai, home to Dubai Healthcare City, the Dhs10 million Center for Innovation and Technology, and ambitious initiatives positioning the UAE as a premier healthcare hub under Vision 2031.

While their paths differ, both cities share a common challenge - translating AI potential into measurable operational improvements. In this session, Paul Swider, Chief AI Officer at RealActivity and leader of Boston's Healthcare Cloud and AI Community, facilitates a collaborative exploration of AI-driven operational efficiency.

This is not a one-way knowledge transfer, it's a dialogue between two innovation ecosystems addressing universal healthcare challenges with complementary strengths.

Building Production-Ready Healthcare AI Agents - From Framework to Operational Efficiency

As the UAE advances its National Strategy for Artificial Intelligence 2031 and Dubai positions itself as a global healthcare innovation hub, regional healthcare organizations are moving beyond simple chatbots to sophisticated AI agents that coordinate complex clinical workflows. But how do you architect these systems for production environments where patient safety, regulatory compliance, and seamless integration with existing EHR/ERP systems are non-negotiable?

In this session, Paul Swider, Chief AI Officer at RealActivity and architect of the Physician Support Copilot, will demonstrate how Microsoft's Agent Framework enables healthcare organizations to build enterprise-grade, multi-agent orchestration systems to improve operational efficiency in healthcare organizations.

Vector Databases, Embeddings, and Semantic Search - The Future of Information Retrieval

This session will explore the role of vector databases, embeddings, and semantic search in modern AI applications. We'll demonstrate how these technologies can be used to create more powerful, intuitive search experiences. We'll also share lessons learned from real-world implementations, focusing on challenges and solutions. This session is ideal for developers interested in the cutting-edge of information retrieval and storing state of LLM applications.

Unleashing the Power of Rust: A Beginner's Guide to Crafting High-Performance Code

If you're new to programming or just looking to expand your skills, Rust is a fantastic language to start with. In this session, we'll cover the basics of Rust programming, including its syntax and essential features. We'll also discuss why Rust is such a unique language, thanks to its memory safety and performance optimizations. By the end of this session, you'll have a solid foundation in Rust and be ready to take your first steps in creating your own API or systems-level programming. So, let's dive in and discover the exciting world of Rust programming together!

Unlocking Insights: How Data Science is Transforming Healthcare

Get ready to be blown away by the latest advancements in healthcare and data! Power Platform, Microsoft 365, Azure, and Dynamics 365 cloud for industry are revolutionizing the healthcare and health life science industry by unlocking powerful insights through data science and engineering. Join us as we dive deep into the Common Data Model, layered industry-specific schemas, mission-critical business app architecture, and multi-cloud data engineering used in Azure Health Data Services and Microsoft Cloud for Healthcare. You won't want to miss this opportunity and discover how these mission-critical data engineering technologies can transform any organization!

Industry Cloud: Microsoft Cloud for Healthcare

Microsoft Cloud for Healthcare runs on Dataverse, Dynamics 365, Microsoft 365, and Azure. Learn how Cloud for Healthcare can assist in optimizing clinical care, patient access, the physician experience, and more. Explore how industry cloud and SaaS services can optimize healthcare organizations.

GraphQL and FHIR APIs

GraphQL is a query language for APIs and a runtime for fulfilling those queries with your existing data. Graphir is a lightweight proxy that reshapes GraphQL queries and mutations into standard FHIR REST requests. In this session, we'll use two open-sourced projects from the Microsoft HLS engineering team to learn how to provide a GraphQL interface over a FHIR API to accelerate client app development.

Rust for Mission-Critical AI: A Journey into Healthcare's Safest and Most Performant Language

Discover the powerful potential of Rust in mission-critical AI applications in this comprehensive 45-minute session tailored for newcomers. Starting with a gentle introduction to Rust, we'll guide you through its key features and benefits, highlighting how Rust can help you create more reliable and efficient IoT, data engineering, and API systems. Delve into the technical aspects that make Rust stand out, such as its memory safety guarantees, fearless concurrency, and strong static typing. Finally, we'll compare and contrast Rust with C# to reveal Rust's unique advantages to mission-critical AI projects. By the end of the session, you'll have a deeper understanding of how Rust can revolutionize your AI systems and help you build a robust and resilient infrastructure for the future.

Journey into Generative AI, Building Your Own AI Copilot

Step into the world of Generative AI with our beginner-friendly session. We'll share how we built our AI Copilot for financial and HR tasks, showcasing the fundamentals of choosing and implementing OpenAI's GPT-4 on Microsoft Azure. From model selection to application development using semantic-kernel, Visual Studio, and VS Code, our journey will provide invaluable insights for GAI developers and newcomers. Join us and turn your curiosity into action as we explore the fascinating real world of the AI copilot project!

We are deploying a private version of ChatGPT for a local hospital. We can demo the process and discuss decisions. We can demo our real development environment.

Semantic Kernel and Agentic Algorithms for .NET Developers

Join Paul Swider for an advanced session on mastering the Semantic Kernel and agentic algorithms in. NET. Dive deep into the Semantic Kernel's architecture, including the Azure.AI.Agents namespace and uncover how to integrate large language models with mission-critical .NET applications.

Learn to create, deploy, and manage powerful native plugins using c#. Additionally, delve into advanced topics such as agentic solutions, intelligent planning, and prompt engineering.

This session is tailored for developers who aim to elevate their .NET skills with advanced AI capabilities. Don’t miss this opportunity to expand your development skills and enhance your expertise.

This session will be a deep dive into the Semantic Kernel SDK. I will focus on the .NET demos. I'll show how the semantic kernel can be used to add conversational capabilities to existing apps, and new apps. Thank you for all that you do!

From Code to Intelligence – Building NextGen AI Agents

You've built functions and optimized workflows but are your AI agents actually intelligent? Building enterprise AI session last year, "Introduction to Semantic Kernel," we now explore semantic-aware functions and advanced agent orchestration.

Explore Agentic Coding tools (such as GitHub Copilot, VS Code, Cursor, and others) to understand how programming is evolving from a purely functional task into a more creative and expressive craft.

In this session, you'll learn:
• Understand Agentic AI Break down the architecture behind intelligent, adaptive AI agents.
• Understand Agentic Coding (Github Copilot, VS Code, Cursor, and others. See how the art of programming rises over the function of programming.
• Orchestrate AI with Semantic-Aware Functions – Use dynamic, context-driven functions to optimize agent orchestration and workflows.
• Manage Semantic App State – Enable AI systems to retain memory, adapt to interactions, and evolve in real-time.
• Apply Microsoft Research Insights – Leverage breakthroughs from AutoGen and Magentic-One to build enterprise-grade AI agents.

Whether you're looking to enhance your current AI applications or build something entirely new, this developer-focused session provides a roadmap to innovate confidently.

Last year's session was great. It was also one of the first live public sessions on the semantic kernel. This year I built on the success.

This session demos cutting-edge Microsoft Research projects, including AutoGen v0.4, Semantic Kernel, and Magentic-One.

Beyond Intelligence—Securing & Scaling Agentic AI (Tokyo Edition)

You’ve built functions and streamlined workflows, but are your AI agents both smart and safe? This developer-focused session dives into semantic-aware functions, advanced agent orchestration, and the security patterns that keep them trustworthy in production.

What You’ll Learn
Decoding Agentic AI – Break down the architecture behind intelligent, adaptive agents and where security controls must reside.

Semantic-Aware Orchestration – Use context-driven functions to optimize workflows while enforcing zero-trust guardrails and policy-as-code.

Managing Semantic App State: Retain memory, adapt in real time, and protect sensitive context from leakage or abuse.

Threat-Model & Defend – Identify common jailbreak vectors, implement prompt shielding, and integrate Azure Confidential Computing & Defender for Cloud into your agent stack.

Whether enhancing existing apps or launching new solutions, you’ll leave with a battle-tested roadmap to innovate confidently and securely.

Building Secure Next-Gen Products with Microsoft AI Developer Services (Tokyo Edition)

Konnichiwa, AI product leaders and solution architects! Join Paul Swider in Tokyo for a deep dive into the secure future of AI product development. This advanced session weaves Microsoft’s newest AI-development stack—Azure AI Studio, Semantic Kernel, and the Healthcare/Enterprise Agent Services.

Key Takeaways:
Secure Agentic Programming: With Azure Confidential Computing + Defender for Cloud, Harden prompts/agents against jailbreaks, data leaks, and supply-chain tampering.

-MCP-Based Authorization at Scale — Orchestrate multi-agent workflows using Model Context Protocol’s fine-grained authorization and policy-as-code (OPA, Azure Policy) for true zero-trust.

-Compliance-Ready Conversations — Build privacy-filtered bots that satisfy APPI and upcoming METI AI guidelines.

Location & Format
Live in Tokyo • Bilingual slides & Q&A • Hands-on code snippets and threat-model templates provided.

Stay ahead. Stay secure. See you in Tokyo!

Radical AI Governance: From JP-Core FHIR to Firewall (Tokyo Edition)

Japan’s health sector is sprinting toward AI-driven care, but every FHIR JP-Core resource you ingest widens the attack surface. In this demo-rich talk, Paul Swider maps a zero-trust reference architecture that merges Azure Health Data Services, Semantic Kernel agents, Confidential Computing, and Responsible AI guardrails.

You’ll learn to threat-model clinical data pipelines, embed bias and lineage checks, and align controls with APPI, HIPAA, and METI’s forthcoming AI guidelines—transforming compliance into a competitive moat while unlocking hyper-personalized patient experiences.

Beyond Intelligence—Securing & Scaling Agentic AI (Lisbon Edition)

You’ve built functions and streamlined workflows, but are your AI agents both smart and safe? This developer-focused session dives into semantic-aware functions, advanced agent orchestration, and the security patterns that keep them trustworthy in production.

What You’ll Learn
Decoding Agentic AI , Break down the architecture behind intelligent, adaptive agents and where security controls must reside.

Semantic-Aware Orchestration – Use context-driven functions to optimize workflows while enforcing zero-trust guardrails and policy-as-code.

Managing Semantic App State: Retain memory, adapt in real time, and protect sensitive context from leakage or abuse.

Threat-Model & Defend – Identify common jailbreak vectors, implement prompt shielding, and integrate Azure Confidential Computing & Defender for Cloud into your agent stack.

Whether enhancing existing apps or launching new solutions, you’ll leave with a battle-tested roadmap to innovate confidently and securely.