What does it actually look like when a company puts AI at the centre of everything, rather than bolting it onto processes that haven't changed in decades?
In this episode, Bruce is joined by Jannik, a Microsoft MVP from Germany who started out in Intune and device management and moved into AI before ChatGPT made it cool. Jannik has built hundreds of AI solutions, including an IT and HR support framework used by 120,000 users. At his current company, he built an agent orchestration suite where every agent gets a CV, a profile picture and an onboarding process, just like any other employee.
They talk about why the biggest model isn't always the right one, how to build "golden questions" to test whether your AI is good enough, and why Jannik thinks you can't build serious AI solutions with Copilot Studio. They also get practical about governance, covering agent identities, sandboxing, conditional access, and how Agent 365 can act as a registry for the agents your users are building.
The thread running through the whole conversation is ownership. The models may be smarter than us, but they still misunderstand, make the wrong call and work from outdated information, so a human needs to stay in charge. The episode ends on a more reflective note: what we risk losing as everything gets faster, more polished and more alike, and why Jannik misses the occasional spelling mistake.
No magic, no hype, just a real conversation about building with AI, one prompt at a time.
Bruce is joined by Nathan Foreman, a principal engineer at Roller. Bruce and Nathan have worked together since their Cookdown days, and Nathan was the one who first showed Bruce how to "write code" with AI around a barbecue in Nashville.
Bruce Cullen introduces the AI for Humanity podcast and shares his personal journey with AI, from his early career in IT support to his current role as a product manager. He discusses the evolution of AI and its impact on various industries, as well as his experiences using AI for design, coding, and product development.