When you build AI on how a team thinks and works, most of the time it gets used. But not always. When it comes down to it, there's work people won't hand over no matter how well the AI is built. It’s worth digging into because it kills more AI efforts than bad tools ever will.

​

In a previous role, I led an effort for the adoption of AI imagery for a creative team. I had a creative lead it, shaping the approach and the prompting. It was supposed to make adoption easy. It didn't. We had team meetings about AI, about the creative possibilities it opened up and I taught the different ways it can help and how to do it. We talked about the objections. Nothing.

​

I knew what the objection was, even though it was never said. “This is ours. Not AI’s.” Crafting images and working with a photographer or selecting stock imagery is a core part of a creative’s job identity. The resistance was never about if the AI was good or that they didn’t understand how to use it. It was about what AI was taking from them.

​

Resistance to AI is usually about ownership, and it shows up different ways. It could look like a lack of training or it can be the common, "AI makes mistakes." These are reasonable things people say when the real reason is "I'm not ready to give this up." Tackling these objections often aren’t the most productive path because you're trying to solve the wrong objection.

​

A lot of adoption advice doesn’t address this. It looks at adoption as an information problem. Explain the tool better. Teach prompting. Explain all the ways it will make work better and easier. I tried all that. Some resistance isn't a gap in understanding. It's that you aimed AI at the wrong thing.

​

There’s some work people are happy to give up. We built a system to support proposal work, work that was slow and full of changes and something that everyone and no one owned at the same time. It was received with open arms, because the system was taking on aspects of the process that the team didn’t want to do in the first place. Use AI on that kind of work and there's little resistance. Just relief.

​

Owned work is different. AI is taking something from people. So the right move is building AI to make the person doing the work better at it.

​

Another example, this time using myself. I was an advertising creative for 20 years. Coming up with ideas was the core of what I did and the part I valued the most. And I built an AI agent to help me. Not to do the work but because I knew that sometimes I got stuck and that I came up with better ideas when I had something to react to. So I built an agentic partner that thought like I thought, challenged ideas and riffed, and gave me 20 different ideas at a time. I worked off the ones that had a spark and I dropped the ones that didn’t. It was always my work and my thinking so I let it help me with the part of the work I loved most. And it made me better at it.

​

To me, that’s the solution. When it’s owned work, the owner needs to choose AI and how to use it. Leaders can set a big vision and give the team the resources, but for adoption to be real and not reluctant, the owners of the work need to shape how AI supports them in the vision.

​

One final thought, and one nobody really wants to talk about. It’s hard to get to any of this by asking. If you ask someone if they'd like an AI partner with the work they own, you probably aren’t going to get at the truth. They'll protect what they own, maybe even unknowingly. So you can’t ask. You can only watch and read their response. The work someone defends is the work they own, and that's all the resistance is telling you. Once you know that, the move's in front of you. Some of it, AI can help them sharpen. Some of it, you leave alone.