Most AI rollouts assume everyone starts in the same place. They rarely do. One person is still asking why AI matters to them, while a colleague already uses it daily.
The AISA framework describes four stages people tend to move through. It draws on change psychology, adult learning and research on self-belief.
How AISA fits the three conditions
AISA describes what happens inside a person. The three conditions in The Human Advantage Programme describe what leaders shape around them.
When people feel safe to try and to say what worries them, they move through the early stages faster. That is Adoption. When people learn to direct and check AI in their real work, skill-building sticks. That is Adaptation.
The four stages
Awareness: "Why should I care?" People test whether AI matters for their own role. Abstract talk about transformation rarely helps. Show AI working on a task they recognise.
Investigation: "Is this safe and reliable?" People test whether outputs can be trusted. They also test whether asking basic questions is safe. Psychological safety is the gate here.
Reich and colleagues found in 2026 that psychological safety predicted whether employees adopted AI tools at all. It mattered most at this early point.
Skill-building: "Let me try this at my own pace." People build confidence through practice. Early results are often disappointing. How leaders frame those disappointments decides whether people keep going.
Forced use before confidence exists produces the look of adoption without the learning. Usage climbs while real capability stays flat.
Adoption: "This is how I work." AI becomes part of everyday work. This comes gradually, through experience, and cannot be rushed.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
The stages are not a straight line
People often move back a stage when a new tool arrives or their role changes. The pattern differs by role and team. Use the stages to understand people, never to grade them.
Spot avoidance early
Some people go through the motions. They use AI when watched and drift back afterwards. Others point to every AI mistake as proof it is not ready.
A simple loop helps: recognise, understand, act.
Recognise the pattern by listening to how people talk about AI. Usage numbers rarely show it.
Understand the cause. It is usually unclear roles, fear of being replaced or too little support.
Act on the cause. Fear of replacement needs an honest conversation about how the role will change. Too little support needs practical help.
What leaders can do at each stage
Awareness. Use specific, role-based examples.
Investigation. Model curiosity. Ask your own basic questions out loud.
Skill-building. Run two-week experiments with a short review at the end. Treat poor results as learning.
Adoption. Protect what is working. Avoid adding new tools faster than people can absorb them.
Try this
Ask your team to take the AI Adoption Stage Assessment. Look at the spread of stages.
Then plan one two-week experiment for people in Skill-building with the Experiment Planner.