What Are the 3 Pillars of Change Management?
The three fundamental pillars of change management are: leadership and vision (setting direction and embodying the change), communication (building buy-in at every stage), and skills development (equipping teams to operate differently). Without these three pillars aligned, no transformation project holds, whether it’s an ERP rollout or an AI integration.
Why These Three Pillars, and Not Others?
Classic models, Kotter with his 8 steps, ADKAR with its five levels of individual change, all converge on the same operational reality: change rarely fails for technical reasons. It fails because leaders didn’t embody the vision, because teams didn’t understand why, or because they were asked to work differently without being given the means to do so.
What I observe in the projects I accompany in Morocco and Europe: technology budgets are often well calibrated. What’s missing is the human structure around them.
Recent signals from Morocco illustrate this tension. Maroc Cloud is deploying Gemini Enterprise, Tata Consultancy Services is betting on its Moroccan subsidiary to serve French-speaking Europe and Rabat’s digital projects, Dyn IT is entering the capital of AI Institute: the tools are arriving fast. The question is no longer “do we have access to the technology?” It’s “are we organized to absorb it?”
Pillar 1: Leadership and Vision
Change cannot be delegated. A CHRO deploying an AI tool in recruitment without the CEO having clearly said “this is our direction” ends up alone facing resistance from middle managers.
Kotter formalized this back in 1996: the first step is creating a sense of urgency, and the second is building a coalition of leaders. Not nominal sponsors. Leaders who talk about the project in executive meetings, who ask about results, who show they’re using the new tools themselves.
In the current context, European rules are catching up with Moroccan players working with Europe, as Le Matin.ma reports. This leadership must therefore incorporate an AI governance dimension. It’s no longer optional. A leader driving AI change must understand the regulatory guardrails, not just the product features.
As I explained in my analysis of Morocco’s AI strategy to 2030, the companies moving fastest are those where the leader has taken a public internal position.
Pillar 2: Communication
Communication in a change project is not a communication plan. It’s an architecture of meaning.
Why we’re changing. What it concretely changes for each team. What stays the same. And above all: what we do with questions that don’t have immediate answers.
The ADKAR model is useful here. It breaks down individual change into five stages: Awareness, Desire, Knowledge, Ability, Reinforcement. Communication intervenes heavily on the first two. If a team member doesn’t understand why the company is integrating an AI tool into their workflow, they will never develop the desire to adopt it.
I’ve built a 6-dimension diagnostic framework to assess exactly where an organization stands on these five levels before launching a deployment. Download the AI Board Pack 2026.
One often-overlooked point: communication must be bidirectional. Field feedback, objections, questions repeated in hallways, these are steering signals. A project that doesn’t capture them is flying blind.
Pillar 3: Skills Development
This is the most underestimated pillar, and the most costly when it fails.
You deploy a tool. You run a two-hour training. And six months later, the vast majority of users have reverted to old habits. Not because they’re resistant. Because the training didn’t create real competence, just superficial familiarity.
In AI integration projects for recruitment, as I analyzed in my article on AI in recruitment in 2026, skills development must be progressive, anchored in the real use cases of each team, and supported over time.
AI culture cannot be decreed. It’s built through repeated use, through deeply trained internal champions, and through an organization that values continuous learning. For leaders who want to structure their own development, this guide on AI training for executives is a concrete starting point.
The Three Pillars Together: A System, Not a Checklist
The classic mistake is treating these three pillars as boxes to check sequentially.
In reality, they function as a system. A leader who communicates well but doesn’t train their teams creates frustration. Excellent training without visible leadership creates enthusiasm without direction. Strong communication without real competence creates anxiety.
Successful projects are those where all three pillars are activated in parallel, with intensities that vary according to the project phase.
If you’re a CEO or CHRO and want to assess where your organization stands on these three dimensions before launching your next AI deployment, request a free diagnostic.
FAQ
What is the difference between change management and project management?
Project management steers deliverables, timelines, and budgets. Change management steers human adoption. A project can be delivered on time and on budget and still be a failure if teams don’t use it. The two disciplines are complementary, not interchangeable.
Is the ADKAR model suited to AI projects?
Yes, with a nuance. ADKAR was designed for classic organizational changes. In AI projects, the Knowledge dimension is more complex because the tool evolves rapidly. Skills development must be conceived as a continuous process, not a one-time event.
How long does change management take for an AI project?
There’s no standard duration. What is certain: projects that allocate less than six months to human accompaniment for a significant AI deployment systematically underestimate resistance and anchoring time. The reinforcement phase, often forgotten, is what determines whether the change holds over time.
Can you manage change without an external consultant?
Yes, if the organization has an experienced internal change management lead and strong leadership. In practice, companies that succeed alone are those that have already lived through several transformation cycles. For a first major AI deployment, an external perspective accelerates diagnosis and avoids blind spots.