1. Introduction
Welcome to another episode recap of Impact of AI Explored! In this edition, we (Gerjon and James) dive into a topic that’s buzzing across tech: AI Agents—what they are, what they aren’t, and why all the hype? We’re joined by Dan Boyles, AKA “The Copilot Guy,” head of AI at First AI Group and a guiding voice on real-world AI adoption. Our conversation winds through agent hype cycles, practical business stories, and how not to lose your head (or your SharePoint) in the rush.
You can find the episode here:
Youtube: https://www.youtube.com/watch?v=XfyFnlrYbFU&
Buzzsprout: https://www.buzzsprout.com/2317233/episodes/17265443
Or your favourite streaming platform.
2. Meet the Guest
Dan Boyles is an AI strategist, podcast host, and currently heads up AI at First AI Group—a company helping businesses globally integrate and scale AI ethically, responsibly, and profitably. Dan is a Microsoft Copilot expert, a keynote speaker, and a leading voice on “human-first” AI strategy. With eight years at Microsoft under his belt and a track record of 50+ successful AI transformations, Dan’s expertise ranges from technical deployments to end-user adoption and governance. He’s also active with the Hello AI Collective and a judge at AI leadership events.linkedin+3
3. Setting the Stage
Why talk “AI agents” now? Everywhere you look—conferences, boardrooms, LinkedIn feeds—2025 is dubbed the year of the agent. But, as we found, there’s both hype and reality. Whether you’re a CTO, developer, or simply AI-curious, understanding what agents actually can do (and what they can’t), is essential for making sense of the evolving workplace. In this episode, expect frank talk on what separates automation from true autonomy, plus clarity on what adoption really looks like, and a few warnings we all need to hear.
4. Episode Highlights
1. “What is a true AI agent?”
Dan breaks it down: most so-called “agents” today are closer to automation—think Copilot Studio, rewrapped Power Virtual Agents, and low/no-code helpers. A real agent, he argues, reasons and acts autonomously, making judgment calls instead of just following workflows.
“A true agent understands what you need, reasons about data, and operates outside strict rules—it gives you what you need, not just what you programmed it to do.”
2. Dealing with Hype and Hurdles
The discussion surfaces an “AI agent hype cycle”—including overblown VC investment, skills gaps, business models doomed for correction, and the stark warning: sort your data out before you try anything fancy!
“If there’s one lesson… it’s: sort your data out. That’s the fundamental thing before you start trying to build anything.”
5. Deep Dive: Groundedness in AI—Why It Matters
At the core of the conversation is a dose of reality: most of today’s agents are really clever automation. Real autonomy (the kind that “gets you to Gatwick by whatever means necessary”) isn’t only a coding challenge; it’s about context, judgment, ethics, and—critically—data maturity. If your SharePoint, AD, or policies are a mess, don’t expect magic from AI.
Dan’s triangle of maturity helps visualize it:
- Base: Individual (no-code, automations for personal bottlenecks)
- Middle: Department (low-code, slightly more complex agents)
- Apex: Organization-wide (real strategic agents, but often out of reach except for enterprises)
The real bottleneck? Clean data, organizational readiness, and a willingness to start small—but start right.
6. Real-Life Stories & Examples
- Onboarding Agent: Dan’s firm built AI agents to automate HR onboarding—handling everything from license procurement to IT access, all without a formal HR department. Result: days saved, less chaos, everyone gets what they need.
- Brand Guidelines Agent: Another favorite: an agent that helps keep outbound branding on point. Upload your graphic, and the agent checks color, font, tone—saving their small marketing team from endless, off-brand adventures.
- Legal AI (and business resistance): In the legal industry, automation could reduce weeks of work to minutes, but not every business wants that—especially billing-by-the-hour law firms.
7. Key Takeaways
- Not all “agents” are created equal: Most are advanced automation, not truly autonomous.
- Start with solid data: If your data estate is a mess, no AI is going to save you—fix it first.
- “ROI” is overhyped—focus on innovation: Efficiency gains from AI are real but show up in soft benefits, creativity, and readiness for future growth.
- Business models (and skills) must adapt: The AI gold rush will see a market correction; only robust, strategically aligned solutions will outlive the hype.
- Security and governance are critical: The next wave is about trustworthy, responsible AI. Expect regulation, breaches, and a scramble for governance solutions.
- Agents will become “invisible apps”: Expect semi-autonomous helpers embedded deeply into daily work (think: modern smartphones).
8. Closing Thoughts
As we wrapped up, we found ourselves repeating a truth echoed in every episode—“Sort your data out!” Only then can you ride (not get swamped by) the next wave of AI. Thanks to Dan Boyles for joining and leveling with us about what’s real and what’s next in AI agents.
What did you think? Do you have agent horror stories or dream use cases? Drop us a comment or tune in to our next episode—more practical AI, fewer hallucinations, guaranteed.
