Most organizations are somewhere in the middle of AI adoption. They are not doing nothing, but they are also not fully transformed. C-Suite has big plans, some employees are using tools informally, teams are running pilots, leaders are trying to create policy, and managers are being asked to support work that is changing faster than the organization can define it.
A lot of the conversation still starts with the technology. Which tools should we buy? Which use cases should we prioritize? How do we measure productivity? How do we control risk? Those are all real questions, but they do not fully explain why AI adoption feels so uneven inside organizations.
My background is as a therapist, so I tend to look at this through a behavior change lens. People usually do not change just because a better tool exists. They change when they understand what is changing, when the next step feels possible, when they have enough support, and when the environment around them makes the new behavior easier to repeat.
That feels relevant to AI adoption because this is not only a tooling shift. For a lot of people, AI changes their relationship to work. It changes what they are good at, what they are expected to learn, how they make decisions, and where they feel confident or exposed. Some people feel excited by that. Some people feel behind. Some people quietly avoid it. Some people are already using AI constantly but have never had a real conversation with their manager about what is appropriate, expected, or risky.
The Transtheoretical Model, developed by James Prochaska and Carlo DiClemente, is often called the stages of change. It was originally used to understand how people move through health behavior change, and it typically includes stages like precontemplation, contemplation, preparation, action, and maintenance. The point is simple but useful: people do not usually move from awareness to sustained change in one step.
That structure helped me think about what is happening with AI inside organizations. A company may not see the issue yet. Then it starts to notice scattered use. Then it prepares through pilots. Then it moves into real adoption. Eventually, if the change sticks, AI starts to reshape work itself.
The model also allows change to be cyclical, which feels important here. AI adoption is not a straight line. A company may move from scattered experimentation to formal pilots to real workflow adoption, but then a new model comes out, a vendor releases a major update, a regulation changes, or a new risk appears. The organization has to reassess. It may need to explore again, pilot again, and adjust the way people work again.
That is the purpose of this model: to make the human side of AI adoption visible enough to act on.