What Are the 4 Steps of Change Management?
Change management rests on 4 fundamental steps: diagnose and create urgency, mobilize and engage teams, deploy and embed new practices, then measure and consolidate results. These steps apply to any organizational transformation, including AI integration into operational processes.
That’s the short answer. Now let’s talk about what it actually means for a leader with a real decision to make.
Why Classic Models Still Hold
Kotter formalized 8 steps in the 1990s. Lewin had 3 in the 1950s. Both are right. But for a CEO or CHRO managing a project today, whether it’s AI adoption, a reorganization, or a merger, four steps are enough to structure action without getting lost in theory.
What I observe with my clients: projects that fail don’t lack technology. They lack human method.
Step 1: Diagnose and Create Urgency
Before moving anything, two questions need answers. Where do we actually stand? And why change now?
Diagnosis is not a comfort audit. It’s an honest reading of the gap between the current situation and what the market demands. In Morocco, several companies are now facing real pressure: players like TCS or local cloud platforms are embedding AI capabilities directly into their offerings. If you don’t move, your competitor will.
Creating urgency doesn’t mean creating panic. It means making visible what is invisible: the hidden costs of inaction, the opportunities passing by, the talent walking out.
As I explained in my analysis of the 3 pillars of change management, this diagnostic phase is often rushed. That’s where everything is decided.
Step 2: Mobilize and Engage Teams
A change project without an internal coalition is a dead project.
This step means identifying the people who will carry the change, not just endure it. Sponsors at the executive committee level. Operational relays within teams. And the skeptics, whom you must understand before you can convince.
In the AI integration projects I work on, this is where the difference between real adoption and surface-level usage plays out. A tool deployed without buy-in quickly becomes unmanaged AI: everyone does what they want, with no coherence, no AI governance.
Communication is not a project deliverable. It’s a continuous process that starts at this step and never stops.
I’ve built a 6-dimension diagnostic framework to assess an organization’s change readiness, particularly around AI. Download the Board Pack AI 2026.
Step 3: Deploy and Embed New Practices
This is the step where most organizations think the work is done. They’re wrong.
Deploying means putting new processes, tools, and accountabilities in place. Embedding means ensuring old habits don’t return the moment project pressure eases.
Concretely, this requires three levers. Process redesign: existing workflows must be rebuilt around new practices, not the other way around. Skills development: training teams is not optional, it’s structural. And updating performance indicators: if you’re still measuring the same things as before, you haven’t actually changed.
In Morocco, initiatives like the deployment of Gemini Enterprise through Maroc Cloud or SME automation by players like AH Digital show that tools arrive fast. The question is not access to technology. It’s the organizational capacity to absorb it.
For more on choosing the right AI tools for your organization, read my decision guide on the best AI for businesses in 2026.
Step 4: Measure and Consolidate Gains
Change is not finished when the project is delivered. It’s finished when new practices have become the norm.
This step requires defining clear indicators from the start. Not activity indicators (number of trainings completed, number of tools deployed), but impact indicators: measurable time savings on a process, error reduction, improved client or employee satisfaction.
It also requires celebrating intermediate wins. Not for hollow internal communication, but to sustain energy over time.
Finally, this step is the moment to capitalize: document what worked, identify what needs adjustment, and prepare for the next wave of change. Because there will be one.
If you’re a CHRO or CEO looking to structure your change management approach, particularly around AI, request a free diagnostic.
What AI Changes in These 4 Steps
AI doesn’t replace the method. It makes it more complex.
It accelerates diagnosis by making visible data that no one was reading. It creates new resistance, because employees fear for their jobs, and that fear is legitimate. It requires AI governance from the start, not at the end of the project. And it makes embedding harder, because tools evolve faster than organizations.
What I see in Morocco, and what Le Matin captures well: the risk is not failing to adopt AI. It’s consuming it without mastering it. Buying a tool is not a strategy. Integrating AI into decision-making processes, with a clear method and governance, is.
FAQ
What is the difference between change management and project management?
Project management drives deliverables: timelines, budget, scope. Change management drives people: adoption, resistance, behaviors. Both are necessary. But confusing the two is the most common mistake in transformation projects.
Is Kotter’s model still applicable in 2026?
Yes, in its fundamental principles. Creating urgency, building a coalition, anchoring change in culture: these human mechanisms haven’t changed. What has changed is the speed at which change must happen and the complexity of the tools being integrated, particularly AI.
How do you integrate AI into a change management approach?
Start with diagnosis: which processes can AI concretely improve? Then train teams before deployment, not after. And define clear AI governance from the start: who decides, who controls, who is accountable for results. AI culture is not decreed, it’s built.
How long does a change management process take?
It depends on the scope of the project and the maturity of the organization. An AI tool integration project in a mid-sized team can take three to six months to reach real adoption. A deep organizational transformation is measured in years. What is certain: moving too fast through steps 1 and 2 always costs more in step 3.