Something revealing is happening at the world’s largest consulting firm. Accenture recently announced that promotion to leadership positions would require demonstrable, regular use of the company’s artificial intelligence (AI) tools — and the organization began tracking individual weekly log-ins to enforce it. The company’s own executives acknowledged the uncomfortable truth behind the policy: persuading senior staff to adopt AI has proven harder than with junior colleagues.
The combination of incentives and accountability may be blunt, but it’s become clear that AI fluency is no longer optional at the leadership level. And Accenture isn’t alone. Microsoft has opened a dedicated global role to build AI skilling programs for C-suite executives, a recognition that even at the very top, the capability gap is real and urgent.
However, while the conversation has shifted to C-suites and entry-level employees, the critical layer in between has been largely overlooked. Middle managers — the people who translate strategy into daily practice — remain underdeveloped, undersupported and underestimated as drivers of AI transformation.
In many ways, this is nothing new. For decades, middle management has been the tipping point between strategy on paper and strategy put into practice. Until organizations fix that, the AI return on investment (ROI) gap will persist. Research shows that 88% of organizations are now experimenting with AI, but 81% report those experiments have yet to deliver meaningful bottom-line results. The tools are there; the ambition is there. What’s missing is the human capability needed to make it work at scale, and that capability lives in middle management.
Why Middle Management Is the Make-or-Break Layer
Middle managers sit at the exact intersection of strategy and execution. They decide whether AI experimentation becomes a cultural norm or a top-down mandate nobody takes seriously. They shape whether teams have the psychological safety to try, fail and iterate. They model the behaviors (e.g., curiosity, adaptability, structured experimentation) that determine whether AI adoption embeds into how work actually gets done, or fades after the next all-hands meeting. Yet currently only 14% of nearly 2,000 managers say they do not face any challenges in “driving effective use of AI across their team.”
L&D is uniquely positioned to equip managers to lead AI adoption across their teams. When middle managers champion AI, it stops being an initiative and becomes a practical reality. When they don’t, transformation stalls … even when leadership is fully committed and the technology is ready.
What L&D Should Prioritize
The risk for L&D is defaulting to training programs that are easy to scale but unlikely to change behavior: tool walkthroughs, AI literacy modules and completion-based eLearning. These initiatives create awareness but don’t drive behavior change.
Effective upskilling for middle managers in the AI era requires investment across four dimensions:
- AI literacy with domain relevance: not theory, but practical application within managers’ specific areas of responsibility
- Workflow redesign skills: the ability to identify where AI genuinely adds value, running structured experiments to test it
- Change leadership: coaching teams through uncertainty, building psychological safety and modeling experimentation
- Business outcome orientation: evaluating AI initiatives through a performance lens, connecting adoption to measurable results.
Format matters, too. Blended learning — combining self-paced digital content with live coaching and cohort-based peer learning — creates space for both skill-building and behavioral practice. It also builds a network of managers working through the same challenges, which sustains momentum beyond the training itself.
Measuring What Truly Matters in Manager Development
Traditional training metrics (i.e., completion rates, satisfaction scores, knowledge checks) tell you whether people showed up. They don’t tell you whether behavior changed.
L&D leaders who connect manager development to observable business outcomes will have a fundamentally different conversation with the C-suite than those who report on learner engagement. That means defining success upfront and tracking a layered set of indicators: the number of AI experiments managers initiate, evidence of workflow redesign in their teams, AI adoption rates among direct reports, and ultimately the business value generated from experiments that get scaled.
This approach requires closer partnership with business stakeholders, but it’s also the only way to credibly demonstrate that leadership development is what’s moving the AI needle.
The Time Is Now
Microsoft is building executive AI skilling infrastructure at the C-suite level. Accenture is using promotion pressure to force adoption at the senior manager level. These are symptoms of the same underlying problem: The human capability to lead AI transformation hasn’t kept pace with the technology itself.
L&D leaders have both the knowledge and the mandate to build a better answer — one grounded in genuine capability development rather than compliance and measured by business impact rather than completion. The organizations that invest in developing their middle management layer now won’t just close the AI ROI gap. They’ll create the conditions in which every other transformation initiative has a chance to succeed.
