Each fall, Wharton’s Human-AI Research Group and GBK Collective release what has become one of the most credible barometers of enterprise AI adoption. Now in its third year, the 2025 study—Accountable Acceleration: GenAI Fast-Tracks Into the Enterprise—captures a moment of transition. AI is no longer the shiny new experiment tucked into a few pilot projects. It has become operational.

According to Wharton’s data, 82% of enterprise leaders now use generative AI at least weekly, and nearly half use it daily. Budgets are rising, optimism remains strong, and three-quarters of respondents already report a positive ROI on their AI investments. (Despite what you may have heard recently from MIT.)

The story coming out of Wharton's research seems straightforward: after two years of exploration and experimentation, AI is finally delivering measurable value. But the most interesting—and perhaps least surprising—finding is this.

“The human side remains the bottleneck—and a key potential accelerant. Morale, change management, and cross-functional coordination remain persistent barriers. Without deliberate role design, coaching, and time to practice, 43% of leaders warn of skill atrophy, even as 89% believe GenAI tools augment work.”

In other words, technology is scaling faster than people can adapt.

Moving from Adoption to Understanding

Wharton’s data points to a paradox that anyone working inside large organizations will recognize. Most companies have mastered the mechanics of AI adoption—tools and governance—but not the human side of AI capability-building. Even though it’s one of the most important factors in successful AI transformation, training investment is down eight percentage points year-over-year, and confidence in training as the path to fluency has fallen by fourteen. (This is wild to me, since training is so critical to success.)

Employees are gaining access to AI faster than they’re gaining time to learn it. And that gap—between adoption and understanding—is where the real problem exists.

The AI-Forward Leader

It’s tempting to see this as a workforce issue, but it’s actually a leadership one. AI doesn’t just need implementation; it needs role design. Wharton’s warning about “deliberate role design, coaching, and time to practice” is less about creating new AI-specific jobs and more about rethinking the ones that already exist.

Every role in an organization now has an AI dimension, and each dimension looks slightly different. For a marketing manager, AI might become a collaborator for research and content development. For an engineer, it might be a debugging partner. For an operations lead, it might be a forecasting tool. The technology’s potential is contextual and role-specific—what matters is that leaders make space for people to explore what augmentation actually looks like in their work.

However, that space rarely exists by accident. It has to be designed.

It Begins at the Top

Wharton’s data shows that organizations moving fastest on adoption share one defining trait: executive ownership. Two-thirds of enterprises now report direct C-suite involvement in AI rollout, and sixty percent have appointed a Chief AI Officer to oversee strategy and accountability.

When leadership treats AI as a core business capability rather than a technical experiment, the rest of the organization follows suit. Conversely, when adoption is left to chance or delegated entirely to middle management, experimentation stalls.

What the report makes clear is that the human side of AI can be both the bottleneck and the accelerant. The same people slowing adoption today can unlock its full potential tomorrow—if they are given structure, support, and permission to learn. This is where leaders matter most.

Designing time for practice and sharing insights gleaned may not show up in a quarterly report, but it will determine which organizations sustain their momentum. Investing in coaching, experimentation, and role redesign isn’t an indulgence; it’s the only scalable path to proficiency. And while Wharton’s findings highlight an emerging skill gap, they also point to a simple truth: capability follows culture.

The real ROI from AI will come from leaders who make deliberate role and skills design a leadership discipline, giving their people the time, context, and confidence to use the technology well.