Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM

Ragnar Heil shares hands-on lessons from building and running autonomous agents with OpenClaw. He compares conversational, agent-first workflows with traditional automation tools, highlights real infrastructure and security trade-offs, and explains why this kind of experimentation accelerates deep understanding of agent architecture. The conversation explores where Copilot fits today, what breaks in early-stage agents, and why learning Linux, scripting, and model plumbing is becoming a core AI skill for practitioners.

👉 Full Show Notes
https://www.microsoftinnovationpodcast.com/816

🎙️ What you’ll learn

• How conversational interfaces like Telegram change the way agents are built and operated
• Where traditional automation adds friction compared to autonomous agents
• Practical infrastructure choices for running experimental agents safely
• Real-world security and governance risks when working with agentic AI
• Why hands-on agent maintenance accelerates architectural understanding

✅ Highlights

• “I literally talk into the microphone and then it builds it overnight without any effort for me.”
• “They’re not really autonomous, most of them are retrieval agents.”
• “Please don’t install OpenClaw on your PC or MacBook.”
• “It could literally delete your whole C drive without your command.”
• “I never learned so much around Linux scripts than in my whole life.”
• “I’m losing my fascination because I have to fix it every single day.”
• “It’s taken me further into understanding agents than I have ever been.”
• “We are in a very, very early stage here.”
• “The interface as we know it is going to go away.”

🧰 Mentioned

• OpenClaw: https://openclaw.ai/ [openclaw.ai]
• Copilot: https://copilot.microsoft.com/
• Microsoft responsible AI standards: https://www.microsoft.com/ai/responsible-ai
• Copilot Studio: https://www.microsoft.com/microsoft-365-copilot/microsoft-copilot-studio/
• Microsoft MVP YouTube Series - How to Become a Microsoft MVP - https://www.youtube.com/playlist?list=PLzf0yupPbVkqdRJDPVE4PtTlm6quDhiu7

✅ Keywords
openclaw, autonomous agents, agentic ai, copilot, copilot studio, power automate, telegram interface, linux scripting, ai security, governance, mcp servers, infrastructure

Microsoft 365 Copilot Adoption is a Microsoft Press book for leaders and consultants. It shows how to identify high-value use cases, set guardrails, enable champions, and measure impact, so Copilot sticks. Practical frameworks, checklists, and metrics you can use this month. Get the book: https://bit.ly/CopilotAdoption

Support the show

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM

This episode features Martin Miller and explores the practical reality of using AI inside organisations. The conversation cuts through hype to focus on where AI genuinely adds leverage, where it breaks down, and why subject matter expertise, data quality, and critical thinking still matter. From AI‑generated code and agent teams to data governance, outages, and deepfakes, the discussion frames AI as a powerful amplifier rather than a replacement. The core message is clear: AI rewards clarity of intent, strong foundations, and human judgement.

👉 Full Show Notes
https://www.microsoftinnovationpodcast.com/815

🎙️ What you’ll learn

• How AI‑generated code fails without clear problem definition
• Why data quality and governance are prerequisites for AI success
• When AI amplifies expertise versus creating costly mistakes
• How organisations misuse AI by chasing headcount reduction
• Why critical thinking weakens when AI replaces core learning

✅ Highlights

• “It’s not about can you write a for loop. It’s about can you define what you’re trying to accomplish.”
• “The more you know, the more you can use this tool.”
• “Nothing like making a decision off of bad data.”
• “AI is a power tool in the box of tools, and it’s a super tool.”
• “It will confidently deliver you results that may or may not be what you need.”
• “Headcount reduction and creating solutions are mutually exclusive.”
• “If you can’t manage your data with or without humans, you need to know that.”
• “You can generate the anomaly of the moment using AI all day long.”
• “There’s no Brent to call when your AI isn’t working.”
• “People that used AI couldn’t remember anything about the paper.”

🧰 Mentioned

• Anthropic: https://www.anthropic.com/
• Claude: https://claude.ai/
• OpenAI: https://openai.com/
• ChatGPT: https://chatgpt.com/
• Gemini: https://gemini.google.com/
• DevOps: https://www.atlassian.com/devops
• The Phoenix Project: https://itrevolution.com/product/the-phoenix-project/
• Foursquare: https://foursquare.com/

✅Keywords
ai strategy, ai coding, data governance, critical thinking, ai agents, devops, site reliability engineering, ai outages, deepfakes, prompt engineering, subject matter expertise, enterprise ai

Microsoft 365 Copilot Adoption is a Microsoft Press book for leaders and consultants. It shows how to identify high-value use cases, set guardrails, enable champions, and measure impact, so Copilot sticks. Practical frameworks, checklists, and metrics you can use this month. Get the book: https://bit.ly/CopilotAdoption

Support the show

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM

Simon Doy and Mark Smith explore the rapid shift from experimentation to production use of AI agents. The conversation focuses on practical agent design, Copilot Studio, RAG quality, and where automation fits better than full agents. Simon shares real client work in the NHS, small business realities, and how partners deliver Copilot adoption and agent-led transformation. A recurring theme is choosing the right level of AI, from automation with AI sprinkles to long running agent workflows, while staying pragmatic about cost, governance, and user experience.

🎙 Full Show Notes
https://www.microsoftinnovationpodcast.com/814

👉 What you’ll learn

• How agentic workflows are moving from demos into real production use
• When simple SharePoint grounding is enough and when Azure AI Search is required
• How to combine automation and agents to avoid slow, interactive AI experiences
• Practical approaches to Copilot adoption for leaders and teams
• Why token usage, orchestration, and model choice are becoming operational concerns

✅ Highlights

• “On Friday, I let Claude Opus go and build me a Teams agent overnight.”
• “The agentic side and being able to have these long running tasks is really shifting things.”
• “Customers are now getting much more open to trying things out.”
• “Copilot Studio is the first port of call.”
• “You don’t get the quality back from the agent that you need with complex content.”
• “How you chunk it is so important for RAG.”
• “Not everything needs to be a full on agent.”
• “Automation with AI sprinkles is sometimes the better option.”
• “My world is fast becoming measured by tokens.”
• “You don’t have to use the most expensive model on everything.”

🧰 Mentioned

• Copilot Studio - https://www.microsoft.com/microsoft-365-copilot/microsoft-copilot-studio
• Azure AI Search - https://azure.microsoft.com/products/ai-services/ai-search
• EmpowerM365 - https://empowerm365.com
• Microsoft MVP YouTube Series - How to Become a Microsoft MVP - https://www.youtube.com/playlist?list=PLzf0yupPbVkqdRJDPVE4PtTlm6quDhiu7

✅ Keywords
ai agents, copilot studio, microsoft 365, automation, rag, azure ai search, sharepoint, claude opus, agentic workflows, copilot adoption, power automate, tokens

Microsoft 365 Copilot Adoption is a Microsoft Press book for leaders and consultants. It shows how to identify high-value use cases, set guardrails, enable champions, and measure impact, so Copilot sticks. Practical frameworks, checklists, and metrics you can use this month. Get the book: https://bit.ly/CopilotAdoption

Support the show

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith