
Magnus Sørensen
Microsoft MVP. Developer. Computer Scientist. Type Validation Advocate
Actions
As a Computer Scientist and Engineer, that loves to improve the way I work, I push the limitations of the xRM SDK daily. My goal as a developer for xRM is to make the developer experience easier for my colleagues. I do this by creating developer tools. Thus, I have gotten the nickname xRM Tooling Wizard.
My goal is to bring best practices from both the academic- and real world into the world of Business Applications - and where else applicable.
Links
From Vibe Coding to Agentic Engineering: Building Production-Ready Dataverse Code with AI
Stop crossing your fingers and hoping AI-generated code works. In this session, we'll move beyond "vibe coding" and into the world of agentic engineering, where AI doesn't just write code that works - it writes code you'd be proud to ship to production.
Since August 2025, I've been successfully deploying AI-written code to real customer environments. The secret? An open-source framework that transforms AI from a coding assistant into a professional developer who writes safe, maintainable, and thoroughly tested Dataverse and Azure code.
What You'll Learn:
* Live demonstration of AI writing production-quality Dataverse plugins, custom APIs, and business logic that looks exactly like senior developer code.
* Deep dive into the framework architecture that enforces coding standards, security patterns, and comprehensive test coverage.
* How agentic engineering differs from typical AI coding: focusing on long-term maintainability, not just immediate functionality.
* Practical strategies to implement this approach in your organization—either by adopting my open-source framework or building your own.
Why This Matters:
While everyone's talking about AI and Copilot, we're actually shipping AI-written code to production. This isn't about prompting tricks or hoping for the best - it's about systematic engineering practices that make AI a reliable member of your development team. The agent even writes more unit tests than most developers (myself included) because it never gets lazy.
Key Takeaways:
* Understanding the shift from ad-hoc AI assistance to systematic agentic engineering.
* Concrete patterns for ensuring AI-generated code meets enterprise standards.
* Access to the open-source framework and implementation guidance.
* Real-world examples from months of production deployments.
Join me to discover how professional Dataverse development with AI isn't just possible - it's already happening, and you can start implementing it tomorrow.
Agentic Engineering for Azure-Dataverse Integration: AI-Built Functions and APIs That Actually Work
Transform your Azure-Dataverse integration development from manual coding marathons into systematic, AI-driven engineering. This session demonstrates how to leverage agentic engineering to build professional Azure Functions and Web APIs that seamlessly integrate with Dataverse - all while maintaining enterprise-grade quality and local testability.
Since August 2025, I've been successfully deploying AI-written Azure integrations to production environments. Moving beyond simple AI code suggestions, I'll showcase my proven approach where AI agents write complete, production-ready Azure Functions and Web APIs. These aren't just snippets that compile - they're fully architected services with proper dependency injection, comprehensive error handling, and extensive test coverage that runs flawlessly on your local machine.
What You'll Learn:
* Live demonstration of AI building complete Azure Functions and Web APIs that integrate with Dataverse, including authentication, data operations, and business logic.
* Architecture patterns that enable true local testing without touching cloud resources.
* How to structure your prompts and frameworks to ensure AI generates properly layered code with separation of concerns.
Why This Matters:
The biggest pain point in Azure-Dataverse development isn't writing the code—it's ensuring it works before deployment. By combining agentic engineering with robust local testing strategies, we eliminate the "deploy and pray" cycle. If you're skeptical that AI can write production-quality integration code, this session will change your mind. The technology and techniques are here today - not in some distant future. With the right framework and approach, AI can generate Azure Functions and APIs that are indistinguishable from those written by senior developers, complete with comprehensive test suites that run entirely on your local machine.
Key Takeaways:
* Complete framework for local development and testing of Azure-Dataverse integrations, battle-tested in real customer environments.
* Patterns for AI-generated code that respects Azure best practices.
* Strategies for mocking Dataverse operations and Azure services in local environments.
* Access to open-source templates and frameworks refined through months of actual use.
Stop debugging in production. Join me to discover how agentic engineering makes Azure-Dataverse integration development faster, safer, and surprisingly enjoyable.
No more secrets! Using Managed Identities in Dataverse Plugins
Most Dataverse deployments have to integrate with Azure at some point. Are you still using the unsafe approach of storing secrets in Dataverse Tables or Dataverse Environment Variables?
If so, this session is for you! I will show you how to use Managed Identities inside Dataverse Plugins to authenticate with Azure endpoints safely, without any secrets.
I will take you along the entire journey. From a new Dataverse and Azure environment to a modern secure integration between them based on plugins. After this session, you will be able to update your code to the latest security standards.
Debug and test plugins, workflows, and security locally!
Want to be able to debug your plugin code and forget all about plugin-traces? Want to be able to create tests which can validate that your implemented business logic works as intended - including what you wrote several years ago? XrmMockup is a test-framework that can simulate your D365 environment locally, along with any plugin and workflow code you have.
I will be going through how to set up XrmMockup, how it works and what exactly it is capable of. With XrmMockup you can perform unit- and integration tests of your D365 pipeline. XrmMockup automatically runs plugins and workflows in the right order, even automatically triggering other plugins and workflows. At the same time, the security model of D365 is enforced just as it would be in your live D365 system.
XrmMockup will provide your tests with an organization service. Using this, your tests can query XrmMockup with any organization request - just like your live D365 system. This means all your business logic is run locally which allows you, as a developer, to set breakpoints and debug your business logic - instead of looking at traces in D365.
I will showcase XrmMockup by implementing some simple business logic using test driven development.
Automated in-memory testing of Power Automate Flows
Tools like FakeXrmEasy and XrmMockup allows a developer to mock a Dataverse environment so their plugins can be run locally.
In this session, I will show how the Power Automate Mock-Up tool can mock Power Automate Flows and together with XrmMockup can test business logic that executes back and forth between Power Automate Flows and plugins.
This adds a new dimension to automated testing in Dataverse that will create more durable applications.

Magnus Sørensen
Microsoft MVP. Developer. Computer Scientist. Type Validation Advocate
Links
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top