AI That Thinks: Dewain Robinson on Agents, Copilot, and What’s Next – HGG677

Posted on by

What happens when AI stops being just a chatbot and starts to reason, plan, and actually do useful work? This week, Dewain Robinson joins me for a conversation on Copilot Studio, AI agents, reasoning models, vibe coding, and the future of software development. We go beyond the usual AI hype and into the practical side, how these tools are already changing workflows, helping with real tasks, and forcing all of us to rethink what productive work looks like. From Microsoft’s internal shift after ChatGPT to real-world examples like managing email, building demo data, handling calendars, and guiding software creation, this one is packed with insight. Thanks for listening!

Audio Only
Apple PodcastsGoogle PodcastsPlayer EmbedShareLeave a ReviewListen in a New WindowDownloadSoundCloudStitcherSubscribe on AndroidSubscribe via RSSSpotifyAudioboomSpreakerOther Options

Why This Episode Matters

A lot of AI conversations still stay close to the surface, which model is best, which tool is fastest, or which subscription is worth paying for. This episode goes deeper.

Dewain and I talk about why the current shift in AI is bigger than better chat tools. The real change is the move into reasoning models, systems that can think through steps, loop, refine answers, and act with more context. That changes the conversation from what AI can generate to what AI can actually do.

If you’ve been experimenting with AI, this episode offers a useful recalibration. There’s still plenty of hype, but there are also very practical ways these tools are already saving time and changing how workflows and software get built.

[00:00] Intro & Spring Check-In
[02:20] Dewain Robinson Returns
[05:50] 21 Years at Microsoft
[10:50] From Bot Framework to Copilot Studio
[13:45] The ChatGPT Turning Point
[20:00] Reasoning Models Explained
[27:00] AI Agents & Real-World Workflows
[35:00] Vibe Coding: What Works (and What Doesn’t)
[45:00] Software Development in the AI Era
[55:00] Real Productivity Gains with AI
[01:08:00] Model Selection & Context Management
[01:14:00] Enterprise AI & Deployment Strategy
[01:18:00] Final Thoughts & Wrap-Up

In This Episode

  • Dewain Robinson returns to talk about his 21-year journey at Microsoft and how that led him into Copilot Studio
  • We discuss the moment ChatGPT shifted priorities and accelerated AI adoption across the industry
  • Dewain explains why reasoning models, not chat alone, are the real breakthrough
  • We explore real-world agent use cases for email, calendar, documents, and workflows
  • We break down vibe coding and why structure matters more than random prompting
  • We debate the future role of software developers in an AI-first world
  • Dewain shares why focused teams and practical rollout may be the best approach for enterprise AI adoption
  • We talk about model selection, memory, and how AI systems may increasingly choose tools automatically

Key Takeaways

  • The real AI breakthrough is reasoning, not just chat-based output
  • AI agents become valuable when they can act, loop, and use tools with context
  • Vibe coding works best when it follows a structured development process
  • Your role shifts from writing code to guiding and validating outcomes
  • Organizations need structure, education, and experimentation for successful AI adoption

Featured Topics

From Bot Framework to Copilot Studio

Dewain walks through his Microsoft journey, from Exchange and Active Directory to IoT, Bot Framework, Power Virtual Agents, and now Copilot Studio. It’s a great look at how one career has tracked with multiple platform shifts.

Why Reasoning Models Matter

The key idea is simple: the breakthrough is not better chat alone. Once AI can think through steps, loop, and refine answers, it becomes far more useful in real work.

AI Agents That Do Real Work

We move past theory and into practice, AI handling email, managing schedules, generating content, preparing demo data, and reducing repetitive work.

Vibe Coding Done Right

Vibe coding can work, but only when it follows a real process: planning, specs, iteration, and validation. The shortcut approach leads to messy results.

Enterprise AI and Practical Rollout

Dewain suggests organizations start with focused teams, learn quickly, and scale AI adoption intentionally instead of trying to transform everything at once.

Expanded Summary

This episode starts with Copilot Studio and quickly opens into a much bigger conversation about how AI is changing work. Dewain brings a practical perspective shaped by years inside Microsoft, which helps explain why this moment feels different from previous tech waves.

The shift to reasoning models and agents is what unlocks real value. Instead of just generating content, AI can now participate in workflows, assist with decisions, and reduce repetitive tasks in meaningful ways. That shift matters for creators, knowledge workers, IT teams, and developers alike.

We also spend time on the cultural side of the transition. Vibe coding sounds playful, but the lesson here is serious: the people who get the most from AI are the ones who pair experimentation with process. Planning, validation, and domain knowledge still matter. In many cases, they matter more than ever.

Dewain’s Microsoft background gives the conversation extra weight. He has lived through several platform transitions already, so when he talks about how quickly priorities changed after ChatGPT, it lands as observation instead of hype. That perspective helps frame where tools like Copilot Studio fit, not as novelties, but as part of a larger move toward AI systems that can coordinate work across tools and data sources.

If you’re trying to figure out what matters in AI right now, this episode is a strong guide. It goes beyond buzzwords and focuses on what these systems actually make possible, where they genuinely help, and what kind of human judgment still separates useful outcomes from flashy demos.

Links

About Dewain Robinson

Dewain Robinson is a longtime Microsoft technologist whose work has spanned Exchange, Active Directory, IoT, Bot Framework, Power Virtual Agents, and now Copilot Studio. That range gives him a useful perspective on this current AI moment, not just as another trend, but as a real platform shift.

Today, Dewain is focused on how AI agents and orchestration can move beyond chat and into practical work. His experience helps connect the big ideas around reasoning models, automation, and enterprise adoption to the day-to-day reality of how these tools actually get used.

Join the Conversation

Full show notes, transcriptions (available on request), audio and video at http://theAverageGuy.tv/hgg677

Join Jim Collison / @jcollison for show #677 of Home Gadget Geeks, brought to you by the Average Guy Network.

WANT TO SUBSCRIBE? http://theAverageGuy.tv/subscribe

Join us for the show live each Thursday at 8pmC/9E at http://theAverageGuy.tv/live

Find Us!

Join us in the Facebook group at https://www.facebook.com/groups/theaverageguy/

On Discord at https://theaverageguy.tv/discord

Get the Home Gadget Geeks subscribe links at http://homegadgetgeeks.com

http://theaverageguy.tv is powered by Maplegrove Partners web hosting. Get secure, reliable, high-speed hosting from people you know and trust. For more information, visit http://maplegrovepartners.com

Popular Tags: Podcast, Home Gadget Geeks, Jim Collison, Dewain Robinson, Microsoft, Copilot Studio, AI agents, reasoning models, vibe coding, software development, enterprise AI, automation, productivity, LLMs, Copilot