Sessions & Speakers
The lineup for Power Conference 2026 is live. Expert talks, live demos, and real-world case studies across Power BI, Microsoft Fabric, Power Apps, Power Automate, and Copilot Studio - plus a dedicated Business stage for adoption, governance, and ROI - all led by MVPs, practitioners, and builders from across the Power Platform community.
Power Apps & Agentic AI - Oh My!
The Microsoft ecosystem is evolving rapidly, and with the rise of agentic experiences, it’s becoming increasingly important to understand which types of agents you can deploy, where they fit best and how to create them. In this workshop, we’ll break down the different types of agents available across the Microsoft stack and explore where they thrive—from Power Apps Canvas Apps to Model-Driven Apps, and beyond. You’ll gain a clear understanding of the strengths and limitations of each approach, helping you make informed design choices depending on your business scenario. We’ll take a hands-on look at Agent Feed, show how to deploy agents into Power Apps, and discuss real-world considerations such as user experience, scalability, and performance. Beyond the technical setup, we’ll also cover the often-overlooked but critical aspects of sustainability, Application Lifecycle Management (ALM), and governance, ensuring your agent solutions remain secure, manageable, and future-proof. Whether you’re building your first agent or refining an existing strategy, this session will equip you with the practical knowledge, patterns, and guardrails needed.
Two Ways to Automate: Agent Flows and Computer Use in Copilot Studio
What A full-day, hands-on workshop that goes deep into two of the most powerful and distinct automation capabilities inside Copilot Studio — Agent Flows and Computer Use. Attendees build the same real-world automation scenario twice, using each approach independently, then analyze the architectural trade-offs, performance characteristics, and governance implications of both. By end of day they have hands-on experience with both capabilities and a clear decision framework for choosing between them on any future project. Why Agent Flows and Computer Use represent fundamentally different philosophies of automation — one structured and API-driven, the other adaptive and UI-driven. Most practitioners have experimented with one or the other, but rarely have they had the structured time and guidance to go deep on both simultaneously. Without that comparative experience, architects and developers make technology choices based on familiarity rather than fit — and that costs projects time, money, and maintainability. This workshop fixes that by giving attendees the rare opportunity to build, compare, and critically evaluate both approaches in a controlled, expert-guided environment. How The day is structured as six progressive lab modules: 1. Architecture overview — how Agent Flows and Computer Use work under the hood, where each fits, and where each breaks down 2. Building the automation scenario with Agent Flows — triggers, conditions, actions, and error handling 3. Extending Agent Flows with advanced patterns — parallel branches, loops, adaptive cards, and human-in-the-loop approvals 4. Rebuilding the same scenario with Computer Use — UI interaction, screen recording, element targeting, and fallback handling 5. Stress-testing both approaches — performance, failure modes, maintainability, and governance implications 6. Decision framework workshop — mapping real-world automation requirements to the right approach using a structured evaluation model Each module combines a structured demo with a hands-on lab, and module 6 uses real scenarios brought by attendees to make the decision framework immediately practical. Who Power Platform developers, RPA specialists, automation architects, and consultants who are evaluating or actively adopting agentic automation capabilities in enterprise environments. Skill level 300-400. Prerequisites: solid working knowledge of Copilot Studio, familiarity with Power Automate cloud flows, basic understanding of UI automation concepts desirable but not mandatory. Top 3 Key Takeaways 1. Deep hands-on experience with both Agent Flows and Computer Use — built on a real automation scenario — giving you the practical confidence to architect and deliver either approach on your next project. 2. A structured decision framework for evaluating any automation requirement against both approaches — covering technical fit, governance risk, maintainability cost, and long-term scalability. 3. A clear understanding of where each approach breaks down in production — failure modes, edge cases, and governance boundaries that only emerge when you push both capabilities beyond the happy path.
Thinking Like A Power Platform Architect workshop
Are you ready to take the leap from maker to architect? Join seasoned Power Platform Architects Neil Rowlands and Keith Atherton as they share their real-world experiences, frameworks, and thinking patterns that empower makers, consultants, and citizen developers to grow into strategic solution architects. This workshop is designed to bridge the gap between building solutions and architecting them. Together, we'll explore the mindset shift required to move beyond app creation and into strategic, scalable, and sustainable platform design. You'll learn how to evaluate requirements with a broader lens, design with enterprise in mind, and balance rapid delivery with long-term governance. What You’ll Learn: * The core principles of Power Platform architecture * How to approach problems like an architect, not just a maker * Navigating the trade-offs between no-code speed and long-term sustainability * Designing solutions with governance, ALM, security, and scalability in mind * Common pitfalls in solution design and how to avoid them * Real-life architectural war stories from enterprise projects Format: * Interactive discussions with architectural decision-making scenarios * Hands-on design challenges and group critique * Live whiteboarding and architecture patterns * Q&A with Neil and Keith for practical guidance and mentorship Who Should Attend: * Power Platform makers ready to level up * Functional consultants seeking architectural awareness * Technical leads and developers moving into architecture roles * Anyone responsible for designing or overseeing enterprise-grade Power Platform solutions Requirements: * A good understanding of Power Platform tools (Power Apps, Power Automate, Dataverse) * Curiosity and a willingness to challenge your current approach
Let's Build a Config‑Driven Automation Engine with Power Automate based on SharePoint Lists
Most teams deal with the same problem. Information comes in, someone needs to decide what to do with it, and it ends up being discussed, reassigned, and tracked manually. That usually means meetings, copy‑paste work, and a constant risk that something gets missed. In this session, I’ll show how to replace that with a config‑driven automation engine built with Power Automate and SharePoint. Instead of manually routing tasks, the system evaluates incoming items, matches them against configurable rules, and automatically sends them to the right destination. To make it real, we’ll walk through a practical example based on Microsoft 365 Message Center messages collected in a SharePoint list. But the approach is universal and can be applied to many other scenarios. Depending on the case, the same item can create a work item in Azure DevOps, be sent to a shared mailbox, or posted to a Teams channel. The logic is not hardcoded. It is managed by teams themselves using SharePoint, without changing the flows. We will also compare this approach with the built‑in Planner integration and explain where it works and where real automation becomes necessary. What you will learn: - how to design a config‑driven routing pattern instead of hardcoding logic in flows - how to separate configuration from execution using SharePoint and Power Automate - how to build a multi‑channel routing mechanism (Azure DevOps, email, Teams) - how to give teams control over automation without breaking governance - how to ensure full visibility and audit of what happens to every item If you are a solution architect, Power Platform developer, Microsoft 365 administrator, or responsible for automation and governance in your organization, this session is for you. By the end, you will understand how to design a reusable automation pattern you can apply to many real‑world scenarios, not just Message Center.
Stop Breaking at Scale: Lessons From Real Enterprise Deployments
Enterprise-scale Power Platform solutions often look simple… until they break under real-world load. In this session, we’ll dive into actual production scenarios where apps, flows, and integrations failed, uncovering why they broke and how to prevent it. Attendees will explore practical strategies, architectural patterns, and anti-patterns to build truly resilient enterprise solutions. We’ll cover event-driven designs, retry policies, multi-system orchestration, SQL backend optimization, and monitoring techniques—all illustrated with real-life enterprise examples. By the end of this session, participants will leave with actionable tactics to ensure high-performance, scalable, and maintainable Power Platform solutions that survive the challenges of enterprise-scale deployments. This is not theory—these are lessons learned at the frontlines of global enterprise projects.
Real AI Use Cases That Matter: From Buzzword to Business Impact
AI has moved beyond prototypes and flashy demos – but many organizations still struggle to turn pilots into production-ready, scalable solutions on the Microsoft stack. In this session, we walk through real AI use cases that actually landed in the real world, built around Power Platform, Azure OpenAI, and Microsoft ecosystem tools. We will dissect several concrete scenarios, including call quality control powered by Power Automate and Azure OpenAI, AI-driven study advice using Dataverse and SharePoint, feedback and review assistants for HR, and process optimization in public sector case handling (WMO) with automatic call transcription and summarization. For each case, we will cover architecture, data flows, prompts, and the Power Platform components involved – and then be honest about what went well, what didn’t, and why. Beyond the technical patterns, we will zoom out to the “AI movement” inside organizations: how to select the right use cases, deal with Shadow AI, set up governance and an AI council, and keep expectations realistic while the hype is raging. You will leave with a pragmatic checklist to evaluate AI ideas, a clearer view of where Power Platform shines for AI scenarios, and concrete patterns you can reuse in your own environment the next day. This session is designed for architects, makers, consultants, and decision-makers who want to connect AI innovation with practical implementation on Power Platform – with real lessons from the field instead of another generic Copilot dem
Your flows are slow. Let's fix that.
Your flows are slow. Let's fix that. Okay so you already know Power Automate. You've built flows. You've automated stuff. You think you're doing alright. But are you doing it fast? This session is for the people who want to stop waiting. Waiting for flows to finish. Waiting for approvals to time out. Waiting for that one connector action that takes forever for no good reason. We're going to fix all of that. I'm going to show you the techniques I actually use to make flows run faster, feel snappier, and handle more without falling over. We're talking parallel branches - running things at the same time instead of one by one. We're talking how to stop over-triggering, how to cut unnecessary HTTP calls, how to use variables smartly so you're not doing the same work twice. And yes, we're going to talk about when Power Automate is simply not the right tool for the job- because sometimes the fastest flow is the one you don't build. This is not a beginner session. I'm not going to explain what a condition is. You know what a condition is. What you might not know is why your flow is slow, where the bottleneck actually is, and what to do about it. Let's go.
Low-Code CI/CD: Tooling, Event Hooks, and Fusion Development
Low-code does not mean no-discipline. The first app is easy. Then there are fifty of them, plus flows and agents, and “export the solution and import it into prod” has quietly become a person clicking through environments on a Friday afternoon. That doesn’t scale, and it doesn’t survive an audit. This session brings real software-engineering discipline to Power Platform without pretending makers are all developers. We’ll cover a sane environment strategy, getting solutions under source control, and automated deployment with Power Platform pipelines and the ALM Accelerator — what to gate, what to automate, and where the manual steps genuinely belong. Crucially, we’ll bring flows and Copilot Studio agents into the same pipeline, so your automation and your agents ship with the same rigor as your apps. You’ll leave with a deployment approach your platform team can actually adopt.
Talking to Apps Power Automate Doesn't Like - on Custom Connectors and Azure Functions
Some apps just don't play nice with Power Automate. Not every app has a built-in Power Automate action. When there isn't one, a custom connector is how you build it - wrapping an external API, an internal API behind an on-premises data gateway, or even simplifying a chatty Graph/REST API call into a single friendly action. Simple API - simple connector. Not so simple API? That's where the fun starts. Non-standard HTTP verb, a literal / inside a path parameter, credentials that can't live in a static connection... That's when a custom connector alone stops being enough, and you need an Azure Function behind it. This session is a tour of both pieces: • how to build and use a custom connector in a flow • how to build and call an Azure Function directly from a flow • how to have a custom connector call an Azure Function instead of the raw API The running example is a real integration to a system Power Automate was never meant to talk with.
From Prompt Flow to Agents: Orchestrating External Systems with OpenAPI Actions & Power Automate
Many teams start with Prompt Flow and need more autonomy and integrations. This session shows how to evolve to Azure AI Foundry Agents with OpenAPI Actions that call line‑of‑business systems (procurement/RFQ, CRM, ticketing). We’ll design Action contracts LLMs handle well, discuss auth patterns, and add a Power Automate flow as a secure relay so browsers don’t hold tokens. We’ll cover idempotency, rate limits, logging, correlation (thread/run IDs), and error handling across boundaries. A practical, governed path from POC to production.
Power Apps & Agentic AI - Oh My!
The Microsoft ecosystem is evolving rapidly, and with the rise of agentic experiences, it’s becoming increasingly important to understand which types of agents you can deploy, where they fit best and how to create them. In this workshop, we’ll break down the different types of agents available across the Microsoft stack and explore where they thrive—from Power Apps Canvas Apps to Model-Driven Apps, and beyond. You’ll gain a clear understanding of the strengths and limitations of each approach, helping you make informed design choices depending on your business scenario. We’ll take a hands-on look at Agent Feed, show how to deploy agents into Power Apps, and discuss real-world considerations such as user experience, scalability, and performance. Beyond the technical setup, we’ll also cover the often-overlooked but critical aspects of sustainability, Application Lifecycle Management (ALM), and governance, ensuring your agent solutions remain secure, manageable, and future-proof. Whether you’re building your first agent or refining an existing strategy, this session will equip you with the practical knowledge, patterns, and guardrails needed.
Thinking Like A Power Platform Architect workshop
Are you ready to take the leap from maker to architect? Join seasoned Power Platform Architects Neil Rowlands and Keith Atherton as they share their real-world experiences, frameworks, and thinking patterns that empower makers, consultants, and citizen developers to grow into strategic solution architects. This workshop is designed to bridge the gap between building solutions and architecting them. Together, we'll explore the mindset shift required to move beyond app creation and into strategic, scalable, and sustainable platform design. You'll learn how to evaluate requirements with a broader lens, design with enterprise in mind, and balance rapid delivery with long-term governance. What You’ll Learn: * The core principles of Power Platform architecture * How to approach problems like an architect, not just a maker * Navigating the trade-offs between no-code speed and long-term sustainability * Designing solutions with governance, ALM, security, and scalability in mind * Common pitfalls in solution design and how to avoid them * Real-life architectural war stories from enterprise projects Format: * Interactive discussions with architectural decision-making scenarios * Hands-on design challenges and group critique * Live whiteboarding and architecture patterns * Q&A with Neil and Keith for practical guidance and mentorship Who Should Attend: * Power Platform makers ready to level up * Functional consultants seeking architectural awareness * Technical leads and developers moving into architecture roles * Anyone responsible for designing or overseeing enterprise-grade Power Platform solutions Requirements: * A good understanding of Power Platform tools (Power Apps, Power Automate, Dataverse) * Curiosity and a willingness to challenge your current approach
Trust by Design: Building Human-in-the-Loop AI Agents with Copilot Studio & Power Platform
Keeping humans in the loop isn’t just a best practice, it’s the difference between trusted AI adoption and operational chaos. As AI agents become more autonomous, designing clear control points for people becomes essential. In this demo-heavy session, you’ll see how to design human-in-the-loop agentic solutions using Microsoft Copilot Studio and the Power Platform, with practical patterns you can apply immediately in real-world environments. This session is aimed at Power Platform makers, Copilot Studio builders, architects, and technical decision-makers who are already familiar with building copilots or cloud flows and want to move beyond simple automation toward responsible, production-ready AI agents. You will learn how to: - Monitor and supervise agent behavior using the Agent Feed to maintain real-time visibility into conversations and decisions. - Design approval and intervention patterns with Advanced Approvals and Request for Information flows, bringing humans back into critical decision points. - Extend Copilot Studio with custom Power Platform solutions to align agent behavior with your organization’s governance and risk model. We will focus on practical design patterns, demos, and architectural trade-offs. We will not cover introductory Copilot Studio concepts, AI theory, or organization-wide adoption strategies. If you want to build AI agents that are powerful, transparent, and accountable, without slowing teams down, this session gives you a clear, proven blueprint.
Spice up Power Apps Development with Claude 🔥
Working with Claude as your coding agent isn’t just a productivity hack, it’s a whole new approach to building and optimising Power Apps. This session showcases the possibilities of implementing Claude as part of your workflow in low code app development. Join MVP Cathrine Bruvold as she demos how to build a Power Apps skill that helps generate YAML code for app components ready to copy+paste, rewrite complex formulas for better readability and debug Power Fx errors that are tedious to resolve. You’ll learn how to utilise a coding agent to design reusable canvas apps components from a single prompt. Understand how to simplify code that impacts maintainability and performance by letting Claude review the app. Gain insights on how you can utilise coding agents to build better Power Apps. Once Claude is part of your development process, going back to manually building apps “alone” feels unthinkable.
Your Data belongs in the new Flow of Work: MCP Apps Deep Dive
Did you ever wish your model-driven app could do more than just sit in a browser tab - and instead show up right where your users are already working, inside Microsoft 365 Copilot? Extending Copilot with your own business logic often means either accepting the limitations of out-of-the-box experiences or diving deep into complex extensibility frameworks that feel out of reach. Not anymore. By extending your app's built-in MCP server with your own custom tools and modern UI widgets, you can shape Copilot conversations that feel less like database queries and more like collaborating with a knowledgeable colleague. Join our session, where we cut through the noise and get straight to what matters 🤓 We'll pick you up on the key concepts, what an App MCP server is, how custom tools work, and where UI widgets fit in, before you join our joyride through a real business scenario end to end: tracking customer case progress, spotting friction in an approval process, and surfacing insights that naturally guide the next action. You'll see how to author a custom tool using the prompt builder designer, generate a polished Fluent UI widget using the /generate-mcp-app-ui skill in Claude Code or GitHub Copilot CLI - more AI-inception than a Christopher Nolan film 🤯 - and deploy the whole thing to Microsoft Teams. No separate infrastructure. No complex pipelines. Just your app, extended. Whether you're a maker who lives in the low-code world, a pro-developer who wants to mix in code where it counts, or a business user who just wants to improve their workflow, this is the session for you 🤝 Key Takeaways Understand the core concepts behind Power Apps App MCP servers, and how custom tools and UI widgets extend what's already built in - and when this approach is the right fit versus other Copilot extensibility options. See a full walkthrough - from authoring a custom tool and generating a theme-aware widget with AI, to deploying your extended app into Microsoft 365 Copilot via Microsoft Teams. Walk away knowing exactly which extensibility path to choose for your next scenario, whether you're a maker, a pro-developer, or bridging both worlds.
Stop Breaking at Scale: Lessons From Real Enterprise Deployments
Enterprise-scale Power Platform solutions often look simple… until they break under real-world load. In this session, we’ll dive into actual production scenarios where apps, flows, and integrations failed, uncovering why they broke and how to prevent it. Attendees will explore practical strategies, architectural patterns, and anti-patterns to build truly resilient enterprise solutions. We’ll cover event-driven designs, retry policies, multi-system orchestration, SQL backend optimization, and monitoring techniques—all illustrated with real-life enterprise examples. By the end of this session, participants will leave with actionable tactics to ensure high-performance, scalable, and maintainable Power Platform solutions that survive the challenges of enterprise-scale deployments. This is not theory—these are lessons learned at the frontlines of global enterprise projects.
Describe It, AI Builds It! Live App Building with Power Apps Generative Pages
What happens when you let AI design your app pages for you? In this fast-paced, interactive session, we’ll use Generative Pages in Power Apps live on stage, driven by prompts from the audience. You’ll see how natural language turns into real, usable app pages in seconds - and what to do when the AI gets a little too creative. Expect live demos, audience voting, and plenty of practical tips for getting the best results from generative experiences without losing control.
From Scratch to Solid: A Reusable Framework for PCF Components
Tired of messy Typescript and inconsistent PCF projects? Want to speed up your delivery, improve code quality and transform your custom projects? Need a standard way of working across all your components? In this session, we’ll show you how we build PCF controls using a robust, reusable architecture that brings clarity and structure to every solution. With a C# developer mindset, this crafted framework tames Typescript's chaos with properties, utilising MobX for intuitive, scalable state management. We’ll dive into how output properties supercharge your controls, enabling dynamic interactions across the form and unlocking powerful, responsive behaviours. By working through a living example and creating code together, this session will enable you for the next wave user interfaces. If you want to build PCF controls that are not just functional but future-proof, this session is your next step.
Low-Code CI/CD: Tooling, Event Hooks, and Fusion Development
Low-code does not mean no-discipline. The first app is easy. Then there are fifty of them, plus flows and agents, and “export the solution and import it into prod” has quietly become a person clicking through environments on a Friday afternoon. That doesn’t scale, and it doesn’t survive an audit. This session brings real software-engineering discipline to Power Platform without pretending makers are all developers. We’ll cover a sane environment strategy, getting solutions under source control, and automated deployment with Power Platform pipelines and the ALM Accelerator — what to gate, what to automate, and where the manual steps genuinely belong. Crucially, we’ll bring flows and Copilot Studio agents into the same pipeline, so your automation and your agents ship with the same rigor as your apps. You’ll leave with a deployment approach your platform team can actually adopt.
From working to Well-Built: Refactoring a real Power App Canvas App
Every Power Apps developer has shipped an app that works — and quietly wished they'd built it differently. In this hands-on session we take a real, multi-screen canvas app — a live event-request form — and refactor it together. You'll turn a copy-pasted progress bar into a reusable component, collapse 25 global variables into one clean record, catch a delegation trap before it bites in production, and run a quick accessibility pass. You leave with the "before" app, the "after" app, and a refactoring checklist you can use on Monday.
RBAC in Dataverse: When “Looks Secure” is a compliance problem
RBAC in Dataverse is often misunderstood. Not because it is complicated, but because it is usually explained through roles and UI settings - instead of real access behavior. Many solutions look secure. Users see only “their” records. Buttons are hidden. Views are filtered. From a compliance perspective, this feels reassuring. Until it isn’t. In this session, we will look at how access control in Dataverse behaves when solutions move beyond demos and into real, multi-user scenarios. We will explore why UI-based access logic creates false confidence, how it breaks under collaboration requirements, and why this gap often turns into a compliance issue rather than a functional bug. Using practical, demo-driven examples, we will focus on Owner Teams and Access Teams as the mechanisms Dataverse provides to model access correctly: ✅ when records should be owned by teams rather than individuals ✅ how to manage dynamic, record-level access without manual sharing ✅ how access changes automatically as users join or leave a team You will learn: 💡 what RBAC in Dataverse actually enforces 💡 why UI logic does not satisfy compliance expectations 💡 how Owner Teams and Access Teams reduce over-permissioning 💡 how to design access models that remain defensible during audits No deep security theory. No “checkbox compliance”. Just practical access patterns that hold up when someone finally asks, “Who can see this data - and why?” (And yes, RBAC sometimes stands for Really Bad Access Control - especially during audits)
Get your MCP and Skills.md here - Agentic Development to the MAX
Buckle up for a full-day workshop that catapults you into the future of AI-driven development for Power BI projects! Discover how GitHub Copilot and MCP servers turbocharge your modeling, reporting, and deployment workflows zipping through tasks at warp speed while uncovering their clever limits and prime sweet spots. Get ready to roll up your sleeves with pulse-pounding, end-to-end exercises that let you wield agentic magic hands-on, emerging as a Power BI wizard armed with killer insights on when AI steals the show and when classic methods hold the fort. **Heads up: This is a fast and deeply technical course on Agentic experiences. Gear up to master cutting-edge tricks that squeeze every drop of awesome from your agentic journeys. Arm yourself with epic prompts to supercharge Power BI and Fabric, blasting through projects faster than a speeding neural net! We'll arm you with plenty of code snippets, a skills.md blueprint, and battle-tested techniques to level up your game starting tomorrow. Unlock a holistic grasp of agentic systems, blending AI smarts with real-world savvy for innovation. Don't miss this AI rush, THIS will be the future of development on Fabric.
Turning insights into action: The art of data communication
Data is everywhere, but how do you turn insights into action? In this full-day workshop, we will focus on both creating reports that matter and exploring how to use them to truly drive decisions. We will start by understanding your audience: who they are, what decisions they need to make, and how to design reports that not only resonate with them but also provide the information they need to make those decisions. Next, we will look at how to create impactful reports, structure dashboards, use DAX for dynamic titles, and apply effective data visualization to help your audience understand the insights. We will then explore how to sell your insights through storytelling techniques and the psychology of communication. It's not just about the data; it's about how you present it to make it matter to your audience - to truly drive decisions. Good presentation and delivery skills are the difference between a report that looks nice and a report that can change the world. Finally, we will cover how to ensure your reports continue to drive action over time, helping users interpret and act on data for lasting impact.
Get Creative with Power BI: Make these core visuals shine!
Core visuals in Power BI are more powerful than most people think. In this session we look at practical techniques to get more out of the visuals you already have, using only what Power BI gives you out of the box. We use icons, font color, conditional formatting, targets, and highlighting not because they look nice, but because they help your users see immediately what needs attention, understand what the numbers mean, and know what to do next. Everything is practical and ready to apply to your own reports.
Empowering our staff: Creating a Power BI training programme
Are you a technical leader that wants to share your expertise with others? Teams in our organisations have both data and business requirements… all we have to do is train our colleagues how to use Power BI Desktop/Power Query and they can rapidly author their own MIBI reporting. By upskilling existing staff, your organisation will gain ownership of the outputs and reduce the need for contractors In this interactive session we will structure a 2-day Power BI training plan outline, giving all attendees an effective framework to use to deliver to their colleagues. I will also share advice from delivering training internally to hundreds of staff over several years at multiple organisations
The Missing Layer in AI for Power BI: Governance, Testing, and Rollback
Most Power BI AI demos focus on speed: generate a measure, explain a model, rename objects, or build something from a prompt. But once an LLM can actually change a semantic model, speed is no longer the main problem. Control is. In this session, we will look at what is missing from most current AI tooling for Power BI: governance, guardrails, rollback, testing, impact analysis, and security-aware workflows. Using SemanticOps MCP as a practical example, I will show how an AI assistant can work with a Power BI semantic model while still respecting rules: which operations are allowed, which require confirmation, what must be tested before deployment, how sensitive data can be masked before it reaches the model context, and how broken changes can be detected or reversed. The goal is not to argue that AI should replace Power BI developers. The goal is to show what needs to exist before AI-assisted development becomes safe enough for real semantic models, real teams, and real production environments. Attendees will leave with a concrete mental model for evaluating AI tools in Power BI: not just “can it generate DAX?”, but “can it be trusted inside a governed development workflow?”
Agentic Apps on Fabric: Build Mind-Blowing Interactive Experiences on Semantic Models
Imagine turning your Fabric semantic models into stunning, interactive apps in minutes. Less Clicking More Building. In this high-energy session, discover the brand-new agentic development experience that lets you choose the agents that do the heavy lifting. No deep coding, no complex frameworks, just point, describe, and watch agents orchestrate everything visuals, custom navigation, writeback, you can have it all directly on your Fabric assets. Live on stage: real product demos that push Fabric to its limits, think no-holds-barred analytics experiences, AI-driven storytelling canvases, and collaborative apps that feel like they were built by a dream team of developers. This isn’t just low-code. This is the future of development inside Microsoft Fabric. You do not want to miss this.
Power Apps & Agentic AI - Oh My!
The Microsoft ecosystem is evolving rapidly, and with the rise of agentic experiences, it’s becoming increasingly important to understand which types of agents you can deploy, where they fit best and how to create them. In this workshop, we’ll break down the different types of agents available across the Microsoft stack and explore where they thrive—from Power Apps Canvas Apps to Model-Driven Apps, and beyond. You’ll gain a clear understanding of the strengths and limitations of each approach, helping you make informed design choices depending on your business scenario. We’ll take a hands-on look at Agent Feed, show how to deploy agents into Power Apps, and discuss real-world considerations such as user experience, scalability, and performance. Beyond the technical setup, we’ll also cover the often-overlooked but critical aspects of sustainability, Application Lifecycle Management (ALM), and governance, ensuring your agent solutions remain secure, manageable, and future-proof. Whether you’re building your first agent or refining an existing strategy, this session will equip you with the practical knowledge, patterns, and guardrails needed.
The Un-Agentic Workshop: Implementing a Solid Foundation for Agentic AI
It’s time to take a step back. It’s time to look at solid foundations, instead of chasing the latest technology trends in a mad rush. And yes, it’s finally time to stop talking about one thing first - “Age…” - ah, nearly slipped up there 😉 In this workshop, That Thing That Must Not Be Named is important. But we need to answer some other questions first. How can we best succeed with adopting this technology, in a scalable and effective way? What do we need to do first in our tenants to prevent an explosion of yet more discarded IT assets? And how can we tie together the various Microsoft cloud technologies to realise this? This workshop is ideal if you’ve moved beyond the hype, and need practical insights and tips on how to rollout That Thing That Must Not Be Named into your organisation. Ideal for both architects and tenant administrators alike, our workshop will breakdown all of the important Power Platform, Microsoft 365, Microsoft Purview, Agent 365 and additional capabilities, and give you blueprint that you will build out during the day that will: * Implement healthy Application Lifecyle Management (ALM) * Protect your organisational data * Provide a detailed overview of everything created in your environments and tenants * Allow you to proactively detect problematic usage patterns * Collect all of the key information points into a unified solution for reporting and analytics By the end of the day, through a mixture of theory, demos and hands-on labs, you can take this blueprint and apply it within your own environment, so that you can confidently grasp the opportunity on offer to modernise, automate and transform your processes. Come prepared to cut through the marketing fluff, to push your skills to the next level and to have some fun with our generative AI “swear jar”, as both presenters and attendees try to avoid saying that word and raise some money in the process for a worthwhile cause.
Two Ways to Automate: Agent Flows and Computer Use in Copilot Studio
What A full-day, hands-on workshop that goes deep into two of the most powerful and distinct automation capabilities inside Copilot Studio — Agent Flows and Computer Use. Attendees build the same real-world automation scenario twice, using each approach independently, then analyze the architectural trade-offs, performance characteristics, and governance implications of both. By end of day they have hands-on experience with both capabilities and a clear decision framework for choosing between them on any future project. Why Agent Flows and Computer Use represent fundamentally different philosophies of automation — one structured and API-driven, the other adaptive and UI-driven. Most practitioners have experimented with one or the other, but rarely have they had the structured time and guidance to go deep on both simultaneously. Without that comparative experience, architects and developers make technology choices based on familiarity rather than fit — and that costs projects time, money, and maintainability. This workshop fixes that by giving attendees the rare opportunity to build, compare, and critically evaluate both approaches in a controlled, expert-guided environment. How The day is structured as six progressive lab modules: 1. Architecture overview — how Agent Flows and Computer Use work under the hood, where each fits, and where each breaks down 2. Building the automation scenario with Agent Flows — triggers, conditions, actions, and error handling 3. Extending Agent Flows with advanced patterns — parallel branches, loops, adaptive cards, and human-in-the-loop approvals 4. Rebuilding the same scenario with Computer Use — UI interaction, screen recording, element targeting, and fallback handling 5. Stress-testing both approaches — performance, failure modes, maintainability, and governance implications 6. Decision framework workshop — mapping real-world automation requirements to the right approach using a structured evaluation model Each module combines a structured demo with a hands-on lab, and module 6 uses real scenarios brought by attendees to make the decision framework immediately practical. Who Power Platform developers, RPA specialists, automation architects, and consultants who are evaluating or actively adopting agentic automation capabilities in enterprise environments. Skill level 300-400. Prerequisites: solid working knowledge of Copilot Studio, familiarity with Power Automate cloud flows, basic understanding of UI automation concepts desirable but not mandatory. Top 3 Key Takeaways 1. Deep hands-on experience with both Agent Flows and Computer Use — built on a real automation scenario — giving you the practical confidence to architect and deliver either approach on your next project. 2. A structured decision framework for evaluating any automation requirement against both approaches — covering technical fit, governance risk, maintainability cost, and long-term scalability. 3. A clear understanding of where each approach breaks down in production — failure modes, edge cases, and governance boundaries that only emerge when you push both capabilities beyond the happy path.
Trust by Design: Building Human-in-the-Loop AI Agents with Copilot Studio & Power Platform
Keeping humans in the loop isn’t just a best practice, it’s the difference between trusted AI adoption and operational chaos. As AI agents become more autonomous, designing clear control points for people becomes essential. In this demo-heavy session, you’ll see how to design human-in-the-loop agentic solutions using Microsoft Copilot Studio and the Power Platform, with practical patterns you can apply immediately in real-world environments. This session is aimed at Power Platform makers, Copilot Studio builders, architects, and technical decision-makers who are already familiar with building copilots or cloud flows and want to move beyond simple automation toward responsible, production-ready AI agents. You will learn how to: - Monitor and supervise agent behavior using the Agent Feed to maintain real-time visibility into conversations and decisions. - Design approval and intervention patterns with Advanced Approvals and Request for Information flows, bringing humans back into critical decision points. - Extend Copilot Studio with custom Power Platform solutions to align agent behavior with your organization’s governance and risk model. We will focus on practical design patterns, demos, and architectural trade-offs. We will not cover introductory Copilot Studio concepts, AI theory, or organization-wide adoption strategies. If you want to build AI agents that are powerful, transparent, and accountable, without slowing teams down, this session gives you a clear, proven blueprint.
Tips and Tricks to Building a Production-Ready Copilot Studio Agent
Building a Copilot Studio agent that works in a demo is straightforward. Building one that holds up when real users ask unexpected questions, when connectors return errors, when your instructions are too vague, and when IT asks how you are going to govern it - that is a different skill entirely. This session is a dense, practical playbook for the production gap. We cover the 10 things that kill agents before they ship: weak instructions that let the model guess, missing auth that breaks in production, no error handling that returns blank responses, a ConversationStart that teaches users nothing, and testing with polished prompts that miss every real edge case. Each tip is paired with a before/after agent instruction snippet or config screenshot. No theory. Everything is something you can apply before your next agent goes live.
Your Data belongs in the new Flow of Work: MCP Apps Deep Dive
Did you ever wish your model-driven app could do more than just sit in a browser tab - and instead show up right where your users are already working, inside Microsoft 365 Copilot? Extending Copilot with your own business logic often means either accepting the limitations of out-of-the-box experiences or diving deep into complex extensibility frameworks that feel out of reach. Not anymore. By extending your app's built-in MCP server with your own custom tools and modern UI widgets, you can shape Copilot conversations that feel less like database queries and more like collaborating with a knowledgeable colleague. Join our session, where we cut through the noise and get straight to what matters 🤓 We'll pick you up on the key concepts, what an App MCP server is, how custom tools work, and where UI widgets fit in, before you join our joyride through a real business scenario end to end: tracking customer case progress, spotting friction in an approval process, and surfacing insights that naturally guide the next action. You'll see how to author a custom tool using the prompt builder designer, generate a polished Fluent UI widget using the /generate-mcp-app-ui skill in Claude Code or GitHub Copilot CLI - more AI-inception than a Christopher Nolan film 🤯 - and deploy the whole thing to Microsoft Teams. No separate infrastructure. No complex pipelines. Just your app, extended. Whether you're a maker who lives in the low-code world, a pro-developer who wants to mix in code where it counts, or a business user who just wants to improve their workflow, this is the session for you 🤝 Key Takeaways Understand the core concepts behind Power Apps App MCP servers, and how custom tools and UI widgets extend what's already built in - and when this approach is the right fit versus other Copilot extensibility options. See a full walkthrough - from authoring a custom tool and generating a theme-aware widget with AI, to deploying your extended app into Microsoft 365 Copilot via Microsoft Teams. Walk away knowing exactly which extensibility path to choose for your next scenario, whether you're a maker, a pro-developer, or bridging both worlds.
Real AI Use Cases That Matter: From Buzzword to Business Impact
AI has moved beyond prototypes and flashy demos – but many organizations still struggle to turn pilots into production-ready, scalable solutions on the Microsoft stack. In this session, we walk through real AI use cases that actually landed in the real world, built around Power Platform, Azure OpenAI, and Microsoft ecosystem tools. We will dissect several concrete scenarios, including call quality control powered by Power Automate and Azure OpenAI, AI-driven study advice using Dataverse and SharePoint, feedback and review assistants for HR, and process optimization in public sector case handling (WMO) with automatic call transcription and summarization. For each case, we will cover architecture, data flows, prompts, and the Power Platform components involved – and then be honest about what went well, what didn’t, and why. Beyond the technical patterns, we will zoom out to the “AI movement” inside organizations: how to select the right use cases, deal with Shadow AI, set up governance and an AI council, and keep expectations realistic while the hype is raging. You will leave with a pragmatic checklist to evaluate AI ideas, a clearer view of where Power Platform shines for AI scenarios, and concrete patterns you can reuse in your own environment the next day. This session is designed for architects, makers, consultants, and decision-makers who want to connect AI innovation with practical implementation on Power Platform – with real lessons from the field instead of another generic Copilot dem
I Will Hack My Agent - The Story of a Copilot Studio Experiment
What happens when you turn the tables on your own AI? In this session, we take you behind the scenes of an autonomous agent built with Copilot Studio - and then trying to break it. You’ll hear the real story of how an agent designed to streamline job applications was stress-tested with prompt injections, tool misuse, and data exfiltration attempts. Discover where the agent held strong, where it failed spectacularly, and what lessons emerged for anyone building AI-driven automation. This is not theory - it’s a candid look at the vulnerabilities, defenses, and design patterns that matter when your agent moves from the lab to the real world. Why Attend? If you’re working with Copilot Studio, Power Automate, or planning to deploy autonomous agents, this session will give you practical insights into: • Common attack surfaces and how to mitigate them • Designing safe, auditable actions and flows • Grounding AI responses to enterprise data • Building resilience without sacrificing usability Join us for a story that blends innovation, security, and a little bit of hacking fun—because the best way to understand your agent is to try breaking it.
Low-Code CI/CD: Tooling, Event Hooks, and Fusion Development
Low-code does not mean no-discipline. The first app is easy. Then there are fifty of them, plus flows and agents, and “export the solution and import it into prod” has quietly become a person clicking through environments on a Friday afternoon. That doesn’t scale, and it doesn’t survive an audit. This session brings real software-engineering discipline to Power Platform without pretending makers are all developers. We’ll cover a sane environment strategy, getting solutions under source control, and automated deployment with Power Platform pipelines and the ALM Accelerator — what to gate, what to automate, and where the manual steps genuinely belong. Crucially, we’ll bring flows and Copilot Studio agents into the same pipeline, so your automation and your agents ship with the same rigor as your apps. You’ll leave with a deployment approach your platform team can actually adopt.
Agents Everywhere: Making Smart Architecture Decisions in the Microsoft Ecosystem
Everyone is talking about agents, but the hardest question is no longer can we build one? It’s what kind of agent should we build, where should it live, and how do we make sure it delivers real business value? As Microsoft continues to expand agentic capabilities across Copilot Studio, Power Platform, Power Apps, Dynamics 365, Microsoft 365, and custom enterprise solutions, organizations are facing a new kind of architectural challenge. The options are growing quickly, but the decision criteria are not always obvious. This session provides a practical decision framework for understanding the agent landscape across the Microsoft ecosystem. Rather than focusing on step-by-step implementation, we’ll explore how to evaluate different agent patterns based on business context, user experience, data access, extensibility, governance, operational support, and long-term maintainability. We’ll look at where agents make sense inside business applications, when they should be embedded into existing user workflows, when a standalone copilot experience is more appropriate, and when a traditional automation or application pattern may still be the better choice. Along the way, we’ll discuss the trade-offs that matter in real enterprise environments: security, ALM, performance, scalability, ownership, sustainability, and adoption. By the end of this session, attendees will have a clearer mental model for navigating Microsoft’s rapidly evolving agent ecosystem and making confident architecture decisions before committing to a specific tool, pattern, or deployment approach. Attendees will learn how to: * Identify the key agent patterns emerging across the Microsoft ecosystem. * Evaluate whether an agent belongs in an app, a copilot experience, a workflow, or a custom solution. * Match agent design choices to business scenarios, user needs, and enterprise constraints. * Recognize the architectural trade-offs behind embedded, conversational, and autonomous experiences. * Apply practical decision criteria around governance, ALM, scalability, security, and sustainability.
Copilot Studio Meets the Command Line: Create Agents at the Speed of a Prompt
WHAT Building Copilot Studio agents has always been a UI-first effort: open the portal, click through menus, drag nodes, configure triggers, test in the canvas, repeat. In this session I'll show you a different way to work — driving the full lifecycle of a Copilot Studio agent from your terminal, in natural language, and pushing the result back into Copilot Studio as YAML. The whole workflow is live on stage, against a realistic agent rather than a hello-world bot. WHY The portal is fine for a single proof of concept. It falls apart the moment a real project arrives: multiple environments, dozens of use cases, several makers working in parallel, and a backlog that needs a feedback loop measured in minutes rather than hours. I've felt this on real client work, and the mix of GitHub Copilot CLI plus the open-source Skills for Copilot Studio plugin from the Copilot Studio CAT team genuinely changes the economics. The plugin packages the YAML schema, design patterns, and anti-patterns the product team uses internally, so what comes out of the prompt is not "valid YAML the model invented" but YAML that follows tested practice. That matters when you're shipping to production, not just demoing. HOW I start from an empty agent created in the maker portal and clone it locally with the Copilot Studio VS Code extension. From there, the work is driven entirely from GitHub Copilot CLI using the three specialised commands the plugin exposes, each tied to a specific phase of the lifecycle. /copilot-studio:author is where the agent grows — generating topics, wiring actions and knowledge sources, scoping variables, and refactoring the architecture when requirements change. I'll demo a refactor of an agent whose topic structure no longer scales with new use cases, which is the situation most of us actually face. /copilot-studio:test takes over once the agent is published, running point tests and batch evaluation suites against the live agent and interpreting the results without leaving the prompt. /copilot-studio:troubleshoot handles the failures every real project hits: wrong topic routing, validation errors, disambiguation problems, and unexpected orchestrator behaviour. Around the demos I'll cover the parts that decide whether this approach actually works in delivery — when CLI authoring beats the portal and when it doesn't, how to keep humans in the loop on generated YAML, source-control hygiene for agent projects, and a sensible split between inner-loop authoring at the terminal and outer-loop validation in the portal. I'll also be honest about the limitations: the plugin is in beta, the Copilot Studio YAML schema can change, and human review of generated output is still essential. Takeaways: You'll leave with a working mental model of YAML-driven Copilot Studio development, hands-on familiarity with the three plugin commands and where each earns its place in the lifecycle, a reusable repository and review pattern for agent projects, a checklist of the gotchas that bite first (auto-update settings, environment selection, schema drift, over-eager refactors), and a realistic view of the productivity gain — and the trade-offs — versus pure portal authoring. If you walk back to your laptop on Tuesday morning and want to try it on a real agent, you'll have everything you need. Tech and tooling: Copilot Studio, GitHub Copilot CLI, the open-source Skills for Copilot Studio plugin, the Copilot Studio VS Code extension, GitHub, and Microsoft 365 / Entra for agent authentication. Claude Code is mentioned as an alternative host for the same plugin. Prerequisites: Working knowledge of Copilot Studio (topics, knowledge, actions), basic comfort with Git, VS Code and a terminal, and a Copilot Studio environment to experiment in afterwards. No prior experience of GitHub Copilot CLI or the plugin is required.
From Prompt Flow to Agents: Orchestrating External Systems with OpenAPI Actions & Power Automate
Many teams start with Prompt Flow and need more autonomy and integrations. This session shows how to evolve to Azure AI Foundry Agents with OpenAPI Actions that call line‑of‑business systems (procurement/RFQ, CRM, ticketing). We’ll design Action contracts LLMs handle well, discuss auth patterns, and add a Power Automate flow as a secure relay so browsers don’t hold tokens. We’ll cover idempotency, rate limits, logging, correlation (thread/run IDs), and error handling across boundaries. A practical, governed path from POC to production.
Low-Code CI/CD: Tooling, Event Hooks, and Fusion Development
Low-code does not mean no-discipline. The first app is easy. Then there are fifty of them, plus flows and agents, and “export the solution and import it into prod” has quietly become a person clicking through environments on a Friday afternoon. That doesn’t scale, and it doesn’t survive an audit. This session brings real software-engineering discipline to Power Platform without pretending makers are all developers. We’ll cover a sane environment strategy, getting solutions under source control, and automated deployment with Power Platform pipelines and the ALM Accelerator — what to gate, what to automate, and where the manual steps genuinely belong. Crucially, we’ll bring flows and Copilot Studio agents into the same pipeline, so your automation and your agents ship with the same rigor as your apps. You’ll leave with a deployment approach your platform team can actually adopt.
Get your MCP and Skills.md here - Agentic Development to the MAX
Buckle up for a full-day workshop that catapults you into the future of AI-driven development for Power BI projects! Discover how GitHub Copilot and MCP servers turbocharge your modeling, reporting, and deployment workflows zipping through tasks at warp speed while uncovering their clever limits and prime sweet spots. Get ready to roll up your sleeves with pulse-pounding, end-to-end exercises that let you wield agentic magic hands-on, emerging as a Power BI wizard armed with killer insights on when AI steals the show and when classic methods hold the fort. **Heads up: This is a fast and deeply technical course on Agentic experiences. Gear up to master cutting-edge tricks that squeeze every drop of awesome from your agentic journeys. Arm yourself with epic prompts to supercharge Power BI and Fabric, blasting through projects faster than a speeding neural net! We'll arm you with plenty of code snippets, a skills.md blueprint, and battle-tested techniques to level up your game starting tomorrow. Unlock a holistic grasp of agentic systems, blending AI smarts with real-world savvy for innovation. Don't miss this AI rush, THIS will be the future of development on Fabric.
Fabric FinOps - Cost Optimization for Fabric
A software-as-a-service offering such as Microsoft Fabric offers a fixed cost per month, making budgeting much easier. That should be the end of the conversation, but it's actually only the beginning - how do you make the most of the capacity units you get with the fixed price? How can you ensure that you are getting the most out of your monthly investment? The different Fabric workloads all have different ways of using capacity units - some are cost CUs per minute of runtime, some cost per query, some are straight forward to calculate, some are not. Take dataflows gen 2 an an example - it used to be the most expensive way by far of doing data integration compared to pipelines or copy jobs, but what does "expensive" mean in the grand scheme of things? Depending on your organization's previous experience and skill level, dataflows might instead end up being the cheapest option! There is much more to optimizing the bang for your buck with Fabric than only looking at the relative costs of capacity units, as it is easy to get lost in the technology instead of looking at the big picture. This session will look at capacity units as a concept, compare how the different workloads drive capacity unit usage, explore ways to optimize workload usage to decrease compute costs, and discuss other factors that impact the overall cost conversation around Fabric.
Delta Table Optimisation - Improving Queries using Delta Partitioning and Liquid Clustering
Data volumes are skyrocketing, and with every new project, the pressure is on for data engineers to deliver performant queries over larger and larger datasets. In this session, we will deep dive into how Delta Lake’s partitioning and Liquid Clustering capabilities can transform query performance in Microsoft Fabric. We’ll be putting these optimisations to the test against a massive dataset to demonstrate real-world impacts on speed and efficiency. We’ll explore the details of Delta partitioning to ensure your data is stored in the most optimal way, reducing query overhead and slashing runtimes. Then we’ll crank it up with Liquid Clustering, an advanced feature that automatically reorganises your data for faster queries. Finally, we’ll show you how to integrate these Delta optimisations into your Microsoft Fabric Lakehouse, so you can power your dashboards, reports, or machine learning pipelines with near real-time insights and without the dreaded performance bottlenecks. By the end of this session, you will: 1. Understand when, why, and how Delta partitioning improves query performance. 2. Enable Liquid Clustering in Delta Lake to keep your data lean, mean, and query-ready. 3. Integrate partitioned and clustered Delta tables seamlessly with Microsoft Fabric for next-level analytics. Brace yourself: This session may contain dangerously optimised partition strategies and an overdose of high-speed query demos! If you’re a data engineer looking for hands-on techniques to crush query latencies and boost productivity in Microsoft Fabric, then this is your must-attend deep dive. Get ready to leave your old, slow queries in the dust.
Agentic Apps on Fabric: Build Mind-Blowing Interactive Experiences on Semantic Models
Imagine turning your Fabric semantic models into stunning, interactive apps in minutes. Less Clicking More Building. In this high-energy session, discover the brand-new agentic development experience that lets you choose the agents that do the heavy lifting. No deep coding, no complex frameworks, just point, describe, and watch agents orchestrate everything visuals, custom navigation, writeback, you can have it all directly on your Fabric assets. Live on stage: real product demos that push Fabric to its limits, think no-holds-barred analytics experiences, AI-driven storytelling canvases, and collaborative apps that feel like they were built by a dream team of developers. This isn’t just low-code. This is the future of development inside Microsoft Fabric. You do not want to miss this.