WHAT Power Automate has been the automation workhorse of the Power Platform for years. But what happens when your process needs to understand intent, classify information, invoke an agent, or make decisions that aren't strictly deterministic? In this session, we'll explore Microsoft Copilot Studio Workflows - the new AI-native workflow engine - and compare it with traditional Power Automate cloud flows. We'll look at where each shines, where each falls short, and how to choose the right tool for the job instead of following the latest hype. WHY Not every business process can be solved with deterministic automation. Some require reasoning, interpretation, or AI-driven decision-making, making Copilot Studio Workflows a compelling alternative. At the same time, not every workflow needs an LLM deciding what to do next - Power Automate often remains the faster, simpler, and more cost-effective choice. We'll compare the real-world trade-offs, including native AI capabilities, licensing, performance, implementation complexity, and maintainability. Sometimes Copilot Studio Workflows can replace several Power Automate actions or external AI services; other times, they simply introduce unnecessary cost and complexity. The trick is knowing the difference. HOW After a quick introduction to the fundamentals, we'll spend most of the session doing what developers actually came for: demos. Lots of demos. We'll share practical solutions with both technologies, compare them side by side, and discuss the architectural decisions behind each implementation. Expect fewer slides, more live interaction, and probably at least one demo that behaves perfectly... because I've sacrificed enough rubber ducks beforehand. TAKEAWAYS After attending this session, you'll be able to: ✅ Understand the strengths and limitations of Copilot Studio Workflows and Power Automate cloud flows. ✅ Recognize when agentic workflows outperform deterministic automation - and when they don't. ✅ Evaluate architectural trade-offs, including AI capabilities, performance, licensing, and maintainability. ✅ Build Copilot Studio Workflows with confidence using practical patterns demonstrated during the live demos. ✅ Make informed decisions based on the problem you're solving, not just the newest button Microsoft added last week.

WHAT Power Automate Desktop has come a long way from being "the thing that clicks buttons when there's no API." With a growing set of agentic and AI capabilities, PAD can now get help building, maintaining, and debugging flows, and even become more intelligent while those flows are running. In this session, we'll explore how agents and AI are changing what's possible with Power Automate Desktop, from AI-assisted flow development and debugging to Copilot Studio agent integration, local LLMs, and computer vision as a fallback for UI automation. WHY Not every process has a nice API waiting for us. Sometimes the only way to automate a legacy application, stubborn web page, or ancient desktop system is still to make the computer do what a human does. And when PAD is the right tool, these new capabilities can make it considerably more powerful - and potentially give it an edge over competing RPA platforms. But with great power comes... yet another way to automate the same thing. As agents, cloud flows, APIs, AI services, and PAD increasingly overlap, knowing how to automate something is no longer enough. We also need to know which tool belongs where. HOW We'll spend just a few slides covering the landscape before diving into a series of practical, demo-heavy scenarios. We'll see agents help create and fix PAD flows, debug problems, and improve existing automations. Then we'll take things further by combining PAD with Copilot Studio agents, local LLMs, and computer vision. Expect less "here's what this feature does" and more "let's see what happens when we actually let AI loose on a desktop flow." TAKEAWAYS After attending this session, you'll be able to: ✅ Understand the latest agentic and AI capabilities coming to Power Automate Desktop. ✅ Use AI to build, maintain, debug, and improve PAD automations. ✅ Combine PAD with Copilot Studio agents, local LLMs, and computer vision to build more capable automations. ✅ Recognize when PAD is the right tool - and when something else should probably get the job. ✅ Identify new ways to make desktop automations more resilient, intelligent, and capable than traditional RPA.