What Is Change Management and How to Build a Culture That Supports It
Résumé du contenu
Change management aligns people, processes, and resources to move organizations from current to future states. It fails when midlevel leaders lack support, short-term wins are absent, and training lags behind rollout. Organizations that embed skills development into daily workflows, rather than treating it as a separate phase, see faster and more durable adoption.
Getting a new tool approved is one challenge. Getting 500 people to actually change how they work is a completely different one. The gap between announcing a change and seeing it show up in everyday workflows is where AI, cloud, and process rollouts stall. Understanding the AI adoption risks before these issues surface is how organizations get ahead of them.
This article breaks down what change management actually involves, why it fails at predictable points, and how to build the kind of leadership development culture where new ways of working stick.
What is change management
Change management is the discipline of aligning people, processes, and resources to move an organization from its current state to a defined future state. It’s not a communications plan or a training rollout. It’s the connective tissue between a decision and the behavioral shifts required to make that decision work.
Three frameworks dominate how organizations approach this discipline. Each one solves a different part of the problem:
| Framework | Focus | Best used when |
| Lewin’s model | Unfreezing habits, transitioning, refreezing new patterns | Understanding why teams resist shifts, even ones they agree with |
| Kotter’s 8-step process | Senior leadership actions, from urgency to cultural anchoring | Executive alignment before scaling a technical initiative |
| ADKAR | Individual-level awareness, desire, knowledge, ability, reinforcement | Diagnosing why one team adopted a new workflow in two weeks while another is still pushing back three months later |
Used together, these change management frameworks help leaders plan for both organizational alignment and individual behavior change, especially when skills development runs alongside communication and process redesign.
None of these frameworks were built for AI adoption specifically. AI initiatives add variables like trust in algorithmic outputs, governance requirements, and the need for AI literacy paths far beyond technical teams.
Why the change management process breaks down
Change efforts fail at predictable points after launch, when ownership gets fuzzy and old priorities return. Naming these failure points early helps leaders plan reinforcement and capability building before adoption stalls.
The most dangerous moment in any organizational change tends to come in the weeks following launch, when initial enthusiasm meets operational reality.
John Kotter’s research on leading change identified patterns that cluster into three categories:
- Leadership gaps: Not establishing urgency, failing to build a guiding coalition, lacking a clear vision.
- Communication failures: Under-communicating direction, not removing obstacles, no short-term wins.
- Sustainability failures: Declaring victory too soon, not anchoring changes in culture.
The midlevel leader problem sits across all three. Senior leaders understand why the change matters and individual contributors get trained on new tools. But directors and team leads, the people translating strategy into daily execution, carry the hardest load without enough support.
Across learning programs, initiatives that invest in midlevel leader capability early see faster adoption rates. This is also where teams resist AI adoption most visibly: managers who weren’t brought into the « why » can’t answer their teams’ questions with conviction.
Build culture through systemic behavioral change
Culture shifts when daily systems change. Adoption improves when workflows, incentives, and manager routines make new behaviors the default path.
Organizations that talk about culture change but only run communications campaigns consistently see behavior revert within weeks. The organizations that sustain change redesign how work gets done. A VP of Engineering who wants to build a culture of continuous learning around AI, for example, doesn’t send emails about the importance of AI literacy programs. Instead, the systemic approach looks like this:
- Sprint planning protects time for skills development rather than treating it as something people do « when they have time. »
- Performance reviews include criteria for applying new capabilities, not just completing courses.
- Managers receive coaching on how to support team members through role-based AI upskilling roadmap paths.
- Cross-functional teams share early wins in weekly standups, making progress visible and repeatable.
This is where instructional design models like ADDIE, SAM, and agile learning design become practical tools. ADDIE’s structured analysis-to-evaluation cycle suits longer change programs with defined milestones. SAM’s rapid prototyping approach fits organizations that need to iterate quickly as AI tools evolve. Agile learning design aligns naturally with sprint-based engineering cultures already used to short cycles and continuous feedback.
Teams that include executive, IT, financial, and HR leaders from the outset, avoid the cross-functional blind spots that stall adoption.
Measure what matters during organizational change
Measure behavior change and not only the activity. The measurement framework that matters separates outputs (training sessions completed) from outcomes (sustained behavior change), and leading indicators from lagging ones.
Leading indicators
Leading indicators answer: are we on track? They include stakeholder engagement rates, whether managers are reinforcing or quietly undermining the change, and early application rates in real work. For AI initiatives, employee trust in algorithmic outputs is a leading indicator most organizations overlook. Tracking it early signals whether governance needs adjustment before resistance hardens.
Lagging indicators
Lagging indicators confirm whether new behaviors are sticking like sustained usage six to twelve months post-implementation, process efficiency gains, and performance improvements tied to the initiative. Tracking upskilling ROI at this level distinguishes program outputs from actual organizational outcomes, and few organizations measure that far downstream.
Close skills gaps to reduce resistance
Skill gaps raise personal and operational risk, which shows up as resistance. Change plans work better when skills development runs in parallel so employees can build confidence and apply new workflows quickly.
Resistance to change is a rational response to being asked to do something new without the skills to do it well. When employees lack capability, they lack confidence; confidence gaps create adoption gaps. Organizations that fail to integrate change management with workforce development from the outset create those barriers.
Tracking AI skills gaps by role, rather than by department, gives leaders a clearer picture of where confidence gaps are most likely to stall adoption. Skills gaps also aren’t always technical. A team lead rolling out AI-powered testing automation also needs to handle pushback conversations, explain the rationale clearly, and coach struggling team members. Role Play simulations give managers a safe space to practice these conversations before they happen live.
How Devoteam achieved 70% workforce AI adoption in three months
Devoteam, a technology consulting company operating across 25 countries, set an ambitious goal: train every employee on generative AI in just three months. The challenge was building consistency across a globally dispersed workforce where roles, skill levels, and day-to-day responsibilities varied significantly.
Using Udemy Business Pro and the GenAI Skills Pack as the foundation, Devoteam ran a global AI upskilling program that covered employees across functions. The program accelerated both design processes and content development across the organization. By the end of the three-month window, 70% of the workforce had completed AI training, and employee attrition dropped by 4%.
What made the timeline achievable was structure. Rather than building a custom curriculum from scratch, Devoteam worked with a ready-made learning foundation that could be deployed globally without lengthy procurement or content development cycles. That’s the dynamic that changes what’s realistic for organizations managing large-scale change: when the infrastructure is already in place, speed of adoption becomes a planning challenge rather than a content challenge.
Teams that need to build AI fundamentals across hundreds of employees don’t have to wait months for a program to be designed before rollout can begin.
Build change-ready teams with Udemy Business
Sustaining organizational change requires ongoing capability building that keeps pace with evolving tools and shifting processes.
Udemy Business connects change management goals to measurable skills development. Leaders can link business objectives to role-specific learning programs across functions from Finance and HR to AI Engineering and Data Analysis. Every course is taught by practitioners solving real problems, which means employees build applicable skills from day one.
Request a demo to see how skills-first learning supports lasting organizational change.
FAQs
What is the difference between change management and project management?
Project management delivers specific outputs within defined parameters: scope, time, and budget. Change management focuses on the human side: shifting behaviors, reducing resistance, and ensuring people sustain new ways of working after implementation ends. Most failed rollouts have solid project management and weak change management.
Which change management framework is best for AI rollouts?
ADKAR tends to be most useful for AI-specific initiatives because it tracks change at the individual level, where AI adoption most commonly stalls. Lewin’s model helps explain why teams resist shifts even when they agree with the direction. Kotter’s 8-step process is most relevant when executive alignment needs to come before anything else. Most organizations benefit from using all three in sequence rather than picking one.
How long does organizational change management typically take?
Most organizations underestimate the timeline. A pilot with 25 users might succeed in weeks, but scaling to 500 introduces new dynamics: cross-functional coordination, competing priorities, and uneven skill levels across teams. Sustained behavioral change, the real measure of success, typically takes six to twelve months to confirm. Programs that treat training as a one-time event rather than an ongoing capability-building effort rarely hold past the initial launch.
How do you reduce employee resistance to change?
Resistance usually reflects a capability gap, not stubbornness. When employees are asked to work differently without the skills to do it well, pushback is a rational response. Three patterns consistently reduce resistance: investing in midlevel leader capability early so managers can answer their teams’ questions with conviction; creating short-term wins that prove the change is working; and running skills development in parallel with rollout rather than treating training as a later phase.