AI Use Policy

Last updated: June 30, 2026

I've spent 30 years helping organizations navigate change, and the last several watching AI become the change that's hardest for most of them to navigate. So it would be odd not to say something about how I use it myself.

This page is my attempt to be specific about  how I am using AI  in my work, where I think it genuinely helps, and where I believe human judgment is still irreplaceable.

What I use AI for

Research and source-finding. AI is useful for identifying where information might exist and surfacing sources I can then go verify directly. I don't treat AI-generated citations as accurate until I've confirmed them myself — confident-sounding hallucinations have burned me, and I've written about exactly that dynamic in the context of AI adoption. Every statistic or study I publish has a verified source URL I can point to.

Drafting and editing. I use AI as a writing assistant to help structure a draft, suggest phrasing, or work through a section that isn't landing. The ideas, frameworks, and professional perspective in everything I publish are mine. AI helps me express them; it doesn't supply them. I've been using an automated grammar support tool for decades, and still leverage suggested tightening or grammar missteps.

Pattern recognition across large amounts of information. When I'm working with a body of research or need to identify themes across many sources, AI is faster than I am at finding patterns. I then apply my own expertise to evaluate whether those patterns are meaningful. I am cautious about the overly optimistic nature of AI partners and carefully consider whether a pattern truly exists or the truth is being stretched too far.

What I don't use AI for

Developing or applying the changecapable method. My method was developed over 30 years of leading organizational change inside large enterprises and four national digital product launches. It comes from direct experience, behavioral science research I've studied deeply, and thousands of hours working with leaders in real change situations. AI didn't develop it and can't apply it. That judgment is mine.

Client insights and stories. Everything I share from client work comes from real experience. I don't use AI to generate illustrative examples or composite stories. If I'm describing something I've seen, I saw it. I often genericize to protect client anonymity, and my recurring composite change leader Cera Day preceded widespread AI usage and serves as an illustrative guide for leaders in real situations.

Professional judgment calls. What to say to a leader who's losing their team. Whether a change plan has the right sequence. How to read a room that's going quiet. These require the kind of contextual human judgment that AI cannot replicate. I'd call it deep expertise versus common intelligence. AI gives you the pattern most sources agree on. Professional expertise tells you when the common pattern is wrong for this situation. Discernment is the ultimate human skill applied to making this judgment call.

Verification. AI can produce something that looks like a fact. I verify it myself. This is non-negotiable for anything I publish. 

A note on human fallibility

Whether I'm working alone or working with AI assistance, I'm human. I make mistakes. I misread research. I write something incomplete or oversimplified. I form an opinion based on the information I have at the time, and new information changes it.

I don't think AI makes that better or worse — it adds a layer of speed and scale that requires its own discipline. What I try to do is be transparent when I've gotten something wrong, update my thinking when the evidence warrants it, and not pretend that working carefully means working perfectly.

If you find an error in something I've published, I want to know.

This page will change

AI tools are evolving quickly, and so is my understanding of how to use them well. What I've written here is accurate as of the date above, and I'll update it as my practice changes: when I start using something new, stop using something, or revise my thinking about where the line should be.

The goal isn't a perfect policy. It's an honest and transparent one.

Questions about how I've used AI in a specific piece of work? Reach out directly.