Mastering AI Transformation Through Project Management

BY Eric J. Sanders, Ph.D. | BY Marc Bara, Ph.D. | 7 MIN READ

Every few years, a new wave of change sweeps through organizations with the same promise: adapt or become obsolete. In 2012, it was agile methodologies. In 2020, digital transformation. Now it is artificial intelligence. Organizations are racing to implement AI transformation initiatives, purchasing enterprise licenses, running prompt engineering workshops, and rebranding every internal process as “powered by AI.”

The urgency feels new, but the pattern is not. A decade of research from BCG, McKinsey and Harvard Business School tells a consistent story: approximately 70 percent of these transformation efforts fail. And the reasons have almost nothing to do with the technology itself. For project managers and organizational leaders, this is both a warning and an opportunity. The question is whether AI transformation will repeat the same cycle, or whether the lessons from past failures can finally break it.

Why AI Transformation Failure Matters for Project Managers

The numbers are sobering. Boston Consulting Group’s analysis of more than 900 digital transformations found that only 30 percent achieved their goals (Forth et al., 2020). McKinsey’s research placed the success rate even lower in traditional industries, between 4 and 11 percent (McKinsey, 2018). These figures predate the current AI wave, but early indicators suggest AI transformation projects are following the same trajectory. As of 2024, roughly 40 percent of AI initiatives still get stuck in the scaling phase, never delivering enterprise-wide value.

Behind every failed transformation are wasted budgets, demoralized teams and lost competitive ground. For professionals pursuing careers in project management and change management, understanding why these initiatives fail is now a core competency, not an optional specialization.

The pattern from past transformations is clear: organizations treated digital change as a technology problem when it was fundamentally an organization development and change management challenge. The gap between buying new tools and actually transforming how an organization works is where most digital transformation failures occur. For anyone watching how companies are approaching AI today, the resemblance is hard to ignore.

What is the Difference Between AI Adoption vs. AI Transformation?

Research reveals a critical distinction that most organizations miss. During the agile transformation era, experts identified the difference between “doing agile” (implementing ceremonies, roles and artifacts) and “being agile,” which meant fundamentally changing mindsets, decision-making structures and organizational culture. The first can be accomplished in months. The second requires years of sustained change management effort.

AI transformation is falling into the same trap. What most organizations call AI transformation today is really AI adoption, including the following:

Purchasing ChatGPT or Copilot licenses
Running prompt workshops
Building innovation labs that produce impressive demos
Measuring success by how many employees “use AI”

None of this constitutes AI transformation. True AI transformation means restructuring decision-making for human-AI collaboration, redesigning work processes from first principles, and rebuilding how expertise and authority flow through an organization. Project management professionals who understand this distinction are essential to bridging the gap between adoption theater and real organizational change.

What Really Works in AI Change Management

The ING Bank case, documented by Harvard Business School (2018), illustrates what successful transformation requires. In 2015, ING, a traditional Dutch bank with 52,000 employees, dismantled its hierarchical structure, eliminated functional departments and reorganized into 350 autonomous squads. The results were significant: development cycles dropped from 18 months to three to six months, mobile app satisfaction rose 20 percent, and ING became the top-rated mobile bank in the Netherlands.

But the crucial detail is the timeline: three years, not three months. ING changed power structures, focused on customer outcomes rather than internal efficiency, built internal capability instead of relying on contractors, and committed to a multi-year change management journey.

BCG found that organizations applying this kind of comprehensive approach to digital transformation achieved success rates of 65 to 80 percent, compared to the 30 percent baseline. McKinsey’s research confirmed that having more than 50 percent of internal employees on the project management team and planning for 24 to 36 months were among the strongest predictors of success. The evidence points in one direction: AI transformation requires sustained project management and change management commitment, not short-term technology deployment.

How to Successfully Lead AI Transformation Projects

Research from Prosci, Kotter and many others converges on practical tips for project managers looking to improve AI implementation outcomes:

Diagnose before prescribing. Audit data quality, clarify what type of AI the organization actually needs (generative AI, machine learning or intelligent automation), and assess integration capabilities before committing resources.
Apply established change management frameworks. Kotter’s 8-step model (Kotter, 1996) and Prosci’s ADKAR model (Hiatt, 2006) are not theoretical suggestions; they are based on studying what separates the 30 percent that succeed from the 70 percent that fail. Organizations with excellent change management are six to seven times more likely to achieve project success.
Plan for years, not quarters. A realistic AI transformation timeline spans 24 to 36 months. Six-month plans with contractor-heavy teams are a documented recipe for failure.
Measure business outcomes, not tool adoption. Counting how many employees “use AI” is the equivalent of measuring agile transformation by the number of stand-up meetings. Focus on customer impact, decision quality and operational results.

Drive Change with Effective Project Management

The evidence is clear: AI transformation failure is not a technology problem. It is a project management and organizational change problem. The organizations that succeeded in past transformation waves did so by addressing the core competencies of effective project management—organizational design, stakeholder alignment, skill development and reinforcement systems.

The good news is that decades of research on organizational development and project management provide a roadmap for what works. The challenge is whether organizations, and the professionals who lead them, will use it.

Next Steps

For those ready to develop the skills needed to lead AI transformation successfully, Elmhurst University’s Master of Project Management (MPM) program provides the foundation in change management, organizational leadership and evidence-based practice that these initiatives demand.

Elmhurst also offers a Graduate Certificate in Project Management, which consists of the first four courses of the MPM curriculum. You can stop there with the certificate, or you can continue with six more classes to complete the full master’s degree.

If you’re ready to lead change in your organization or position yourself for a new role, fill out the form below to learn more about the Elmhurst MPM program.