Part 1: Nobody Decided Who Decides — The Real Reason AI Adoption Stalls

The Real Reason AI Adoption Stalls

If you've spent any time in AI strategy conversations lately, you've probably heard some version of this: business teams should own how AI gets used in their workflows, and IT should own how it's controlled. It sounds reasonable, and on paper it even sounds like a plan; but it's not a decision right, it's a value statement, and the difference between the two is exactly where most organizations are getting stuck.

A quick distinction worth making

There are many different jobs in any organizational change, but lets discuss two main jobs: designing the structure, and helping people move through it. Think of it like building a building. An architect / engineer decides where the load-bearing walls go, how the rooms connect, and who has access to what; that's structural work, done before anyone moves in. Then there's the move-in team, the people who help everyone settle, find their desks, and learn the new layout. Different job, different timing, different skill.

An OE advisor sits on the architect / engineer's team, helping leaders design that structure. An OCM practitioner sits on the move-in team, helping people adapt once the structure exists. Both teams have other roles too, but those two are worth naming here, because this article is about the architecture, specifically about who has the right to decide things when it comes to AI, not about how to help people adjust once that decision has already been made.

The pattern hiding behind two "opposite" problems

Talk to enough leaders and you'll hear two complaints that sound unrelated. One is "we can't get AI initiatives approved, everything stalls in committee," which is the bottleneck; the other is "people are just using whatever AI tools they want, and we have no idea what's actually happening," which is shadow AI. These look like opposite problems, since one org is too slow and the other too loose, but they trace back to the same root cause: nobody has a clearly assigned right to say yes or no on a given AI use case.

When that right is missing, you end up with one of two outcomes. Either people are afraid to act without permission they can't find, so things stall; or people just act because no one ever told them they couldn't, so things sprawl. It's the same coin, just two different sides.

Principles aren't decision rights

Most AI governance documents are full of principles: use AI responsibly, keep a human in the loop, protect data security. These are fine values to have, but a principle tells you what an organization believes; it doesn't tell you who gets to decide, on what, or how fast. Those are decision rights, and most organizations have written the first thing without ever getting around to the second.

A simple test

Here's a quick gut check you can run today. Pick any AI use case in your organization, big or small, and ask yourself who can say yes to this, and who can say no. Not a department, not a committee, but an actual person who currently sits in that role and could answer the question right now if you called them. If you can name that person in one sentence, you're in decent shape, because that means the right has actually been assigned and someone knows they own it. But if the honest answer is "IT" or "the business" or some other group name, that's not a decision-maker, that's a deflection, and it usually means the call gets made by whoever happens to be in the room that day. And if it takes a paragraph, a meeting, or a shrug just to figure out who that person even is, that's your real gap, and it's not a training gap or a culture gap, it's a structural one.

That gap is exactly what's producing both the bottlenecks and the shadow AI you're seeing right now, and it's worth sitting with that for a minute, because the fix isn't more training or a stricter policy memo, it's naming who actually holds the call.

Next time: why this gap gets a lot more urgent once AI agents start acting on their own, and what to actually do about it.

This is Part 1 of a two-part article. Read Part 2 here: Part 2: Nobody Decided Who Decides — Why It Gets Worse With Agentic AI

Contact

Reach out for consulting or speaking inquiries

Email

hello -@- unbrokenwork -dot- com

© 2026. All rights reserved.