Change models are tools. Each captures something real about how change works, and each has limits. The skill lies in choosing and adapting them.
Change is hard to land. Treat the widely quoted claim that 70% of change programmes fail with caution, but the disappointment is real.
McKinsey's State of Organizations 2026 found 88% of organisations experimenting with AI. Yet 81% report no meaningful bottom-line gains.
Lewin: unfreeze, change, refreeze
Kurt Lewin's model has three stages. Unfreezing builds the motivation to change. Changing develops new behaviours. Refreezing embeds them as the new norm.
Strength. It shows that change needs a reason before a plan. Skip unfreezing and people drift back to old habits.
Limit. Real change is rarely linear. Resistance appears throughout, and people move at different speeds.
Kotter: eight stages
John Kotter set out eight stages. Create urgency, build a coalition, form and share a vision, empower action, win early, consolidate gains and embed change.
Strength. It attends to politics and momentum: credible sponsors, early wins and cultural embedding. Kotter argued change often stalls when leaders skip urgency or coalition building.
Limit. It assumes a planned change with a clear destination and senior alignment.
ADKAR: change one person at a time
Prosci's ADKAR model tracks five individual stages: Awareness, Desire, Knowledge, Ability and Reinforcement.
Strength. It works as a diagnostic. If change stalls, ask where: do people understand why, want it, know how, or have the ability? Each gap needs a different response.
Limit. Individual readiness falls short when structures, culture or leaders work against the change.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
Technical or adaptive?
Ronald Heifetz distinguishes technical problems from adaptive challenges. Technical problems have known solutions and need good planning. Adaptive challenges need people to learn new ways of thinking and working.
Most change mixes both. The common mistake is treating an adaptive challenge as technical: writing a full plan and executing it while the answer is still unknown.
Adaptive work calls for small experiments, fast feedback and adjustment. Use structured models for the technical parts.
AI change is mostly adaptive
AI adoption has no settled destination. Tools change monthly, and the best uses emerge from the people doing the work.
Microsoft's 2026 Work Trend Index found organisational factors accounted for 67% of the AI impact people reported. Individual factors accounted for 32%.
Deloitte's 2026 Global Human Capital Trends found organisations focused on work design were twice as likely to exceed their AI return expectations.
Stopping matters as much as starting. McKinsey found 43% of leaders say they divested assets too late, or failed to. Spreading what works and stopping what does not keeps the organisation ready as change speeds up. That is Acceleration.
Try this
For a change you are leading, list the groups who need to move. Use ADKAR to find each group's biggest gap. Design one specific action for that gap.
Then take the free Change Readiness Scan. For leaders' own part in change, see our change management training.