AI can spin up a slick prototype in seconds, but a demo that looks ready and a system that actually ships and gets adopted are two very different things. In this episode of Domo’s podcast, Governed Data for Agentic AI, Chris Willis, Chief Design Officer at Domo, speaks with Dan Hendriksen, who leads Domo’s Advanced Customer Enablement (ACE) program and spends his days helping customers turn AI experiments into things they can actually run. He’s watched the same patterns play out across build after build, and he’s here to share what separates the projects that ship from the ones that stall.

This conversation is all about “plausibility theater”: the trap of building something that looks like it should work but quietly skips everything real systems need (user volume, concurrency, security, edge cases). Dan explains why the planning conversation matters more than ever now that AI has taken the developer out of the loop, why “do not code” and plan mode are becoming essential habits, and why the time you save on a prototype is best spent up front with stakeholders instead of iterating 45 times on the wrong thing.

Chris and Dan also get into the bigger shifts AI is setting off: the rise of curious generalists over stay-in-your-lane specialists, the long tail of low-value problems that’s finally worth solving, and the move toward bespoke apps that do exactly what a team needs. Along the way Dan makes the case that some of the best AI is maybe not AI at all, and that it all starts with a deceptively simple question: Why? Listen for a practical, field-tested take on getting past the demo and building AI that ships and works.

After listening, you'll learn:

• What “plausibility theater” is and how to avoid it

• Why the planning conversation beats jumping straight to code

• How roles are shifting toward curious generalists

• When the right answer is a simple automation, not an app