AI and the Future of Work: Artificial Intelligence in the Workplace, Business, Ethics, HR, and IT for AI Enthusiasts, Leaders and Academics

385: From API Management to Agent Control: Why Governing AI Actions Is the Only Path to Enterprise Value, with Oren Michels, Co-Founder and CEO of Barndoor AI

Apr 20, 2026

Guest

Oren Michels

Oren Michels, entrepreneur and CEO of Barndoor AI who built Mashery and produces Broadway shows, talks about governing agentic AI. He explains why agent control differs from API security. He describes least-privilege, tool-limiting, and auditing approaches. He warns about BYO-AI risks and how gradual, monitored rollouts can unlock real enterprise value.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Agentic AI Is Defined By Its Actions

Agentic AI must be allowed to take actions, not just suggest them, to unlock transformative value in enterprises.
Oren Michels argues an agent only becomes real when it writes or acts, shifting the problem from advising to controlling behavior.

INSIGHT

AI Trust Needs Controls Beyond Identity

Trust for AI requires different guardrails than for humans because AIs lack life experience and fear of consequences.
Oren Michels compares hiring controls for humans to programmatic controls needed for agents that won't self-police.

ANECDOTE

From Mashery To Barndoor History Repeats

Oren founded Mashery because businesspeople needed to manage APIs without engineers acting as gatekeepers.
That history repeated: companies now need a control plane so non-engineers can safely govern agents.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Send us Fan Mail

Oren Michels is an entrepreneur, investor, board member, and advisor to technology startups in the US and Europe. He is the co-founder and CEO of Barndoor AI, the control plane for agentic AI, and the founder who previously helped define the API management category with Mashery, acquired by Intel in 2013. He is also a Tony-nominated Broadway and Off-Broadway producer whose credits include Romeo+Juliet and Good Night, and Good Luck starring George Clooney.

In this episode, Oren draws on two decades of building foundational infrastructure for the enterprise to make the case that governing AI agents is not a security problem. It is an entirely new category of problem, and most companies do not yet have the vocabulary to describe it, let alone the tools to solve it.

If your agents can already write to your CRM, interpret your instructions, and act without life experience or fear of consequences, who is actually in control?

In this conversation, we discuss:

Why securing AI agents is entirely different from managing APIs, and why traditional security and identity access tools were never designed to handle what agents can do.
The reason most so-called agentic AI is still glorified robotic process automation, and what it will actually take to unlock enterprise value.
How Barndoor AI's "least privilege" framework for agents works, and why the permission logic goes far beyond the identity of the human using the tool.
Why an agent with delete access to your CRM is one probabilistic misfire away from a catastrophic outcome, and why ultimate responsibility always comes back to the humans operating the tools.
The BYO AI parallel to BYOD: why well-meaning employees using personal AI tools with company data may force the enterprise governance moment no one is ready for.
Why the same instinct that took Oren from API infrastructure to Broadway and back to enterprise AI may be exactly the mindset the agentic era demands from its builders.

Explore this conversation:

00:00 Intro and fun Fact

03:46 Oren Michels's Path From API Management to Building Barndoor AI

05:44 Redefining Trust: AI Lacks Life Experience and Fear of Consequences

08:24 History Repeating: Why AI Needs a Control Plane Just Like APIs Did

12:35 Deterministic APIs vs. Probabilistic Agents: Why Governing AI Is a Social Challenge

18:25 How Barndoor AI's "Least Privilege" Framework for Agents Actually Works

20:50 The Token Economy and Context Windows: Wandering Into the AI Home Depot

25:25 Preventing Catastrophic Failures: Why AI Agents Should Never Have Delete Access

31:39 The BlackBerry Moment of AI: Navigating the "BYO AI" Enterprise Trend

38:04 Balancing Tech and Creativity: From Enterprise AI to Producing on Broadway

Resources