Why AI Fluency Is a Leadership Problem

Enterprise spending on generative artificial intelligence has passed $30 billion worldwide, and 95% of organizations said they haven't seen a return on that spend, according to MIT's Project NANDA. The tools are in place. The licenses are purchased. But the productivity pay-off promised by AI keeps missing the mark.

The instinct in many HR and learning organizations has been to close the gap with more training. U.S. businesses spent $102.8 billion on training in the workplace in 2025 alone, and worldwide, global training spend has reached $366 billion a year. But only 15% of learning and development (L&D) leaders said their organizations’ leadership programs are highly effective. The result of trying to teach AI literacy as a skill to be trained on is something we’ve seen time and again: awareness, but not adoption.

Jamie Albers, a member of the 2026 SHRM Labs WorkplaceTech Accelerator and co-founder and co-CEO of Mento, said she sees it differently: Companies may be investing millions into AI technology, but they’re failing to invest in the people who make it useful: the manager layer.

What Managers Are Actually Saying

More than 90% of coaches in Mento reported that AI-related topics have come up in coaching, with 83% of those topics reflecting fears of falling behind, 73% relating to job security, 73% concerning AI adoption pressures, and 53% being about overload and exhaustion. However, qualitative survey data from Mento's members showed that only 2% of managers talk explicitly about AI tools.

That tells us AI is permeating the emotional context of work, as well as the leadership context, but it's showing up as something other than a tool-training problem.

"Managers aren't primarily saying, 'I don't know how to use ChatGPT,' " Albers said. "They're saying, 'I'm overwhelmed, I'm uncertain, and I'm not sure how to lead through this much change.' "

The data confirms this trend. Forty-five percent of middle managers reported experiencing burnout, more than any other group, per data from Future Forum and Gallup. The average manager oversees 12.1 direct reports, up from 10.9 last year; some manage more than 25. In Mento’s own pulse survey of upcoming participants, nearly half of the incoming members rated their ability to stay productive without burning out with a six or below (out of 10). While AI is spurring demands for speed and innovation, the group taking the heat is the manager layer, already stretched the thinnest.

How the Manager's Job Changes

The assumption seems to be that AI will make management easier by lessening the burden of tedious execution. But actually, in this period of transition, AI makes management more difficult, at least temporarily. The problem with AI is that it expands people's capabilities. As a result, we generate more ideas, more possibilities, and more pressure to know where to go.

That changes the shape of the work. Managers used to help teams move from point A to point B. Now, they're leading builders whose paths keep shifting underneath them. An engineer can prototype a design. A marketer can build workflows. A designer can ship code. Cross-functional overlap is up. Laying out clear outcomes, removing roadblocks, and translating ambiguity into priority has gotten harder, not easier.

"Managers are leading in motion now," Albers said. "They're setting direction knowing it may change and helping people build capabilities that didn't exist six months ago."

The Capabilities That Matter Now

The leadership skills that become increasingly important in an AI-saturated world are those that enable teams to function within ambiguity. That includes judgment, prioritization, delegation, resilience, and the ability to lead what Mento calls "builders." A builder goes beyond “doing tasks.” A builder tries things, prototypes, and adapts. More roles are becoming “builder” roles because AI spreads capabilities such as data analysis, design, research, and prototyping to people who didn’t have them before.

What’s becoming obsolete is a management style premised on static execution: planning things out, assigning tasks, and tracking their completion. That approach has its place, but it’s no longer enough.

"AI fluency at the leadership level is about more than knowing which tools to use," Albers said. "It has three parts: cultivating builder mindsets, creating a continuous learning environment, and empowering teams to judge where effort will actually create value."

Why the Old Training Model Falls Short

Leadership development as we knew it was predicated on relatively stable contexts. You send a manager to a two-day workshop, give them a few frameworks, then send them back to work. The limitations of this model are well documented. Most of the learning from one-off training evaporates just weeks after it occurs; less than a quarter of businesses said their leadership development programs are current.

This becomes more pronounced in an AI-rich environment. There’s no end point to converge. According to Mento’s coaching data, the common denominator of successful managers isn’t that they’ve completed a superior training course, it’s that they have a different relationship to learning.

"The people furthest along usually aren't the ones who attended the best workshop," Albers said. "They're the ones experimenting, learning from peers, sharing what works, and staying willing to feel uncomfortable while they build new habits."

What Successful Leaders Are Doing Differently

Mento has noticed a consistent pattern among managers leading their teams well through this period. Five behaviors stand out:

1. Modeling learning in public. The best leaders aren't those who pretend to be the "AI expert" with every tool; rather, they ask questions, get into the experiments in front of others, and set the expectation that it is OK for the most senior leaders to be learning a skill.
2. Turning experimentation into collective learning. Left unaided, adoption of AI lingers at just a few power users. Great managers implement structures for teams to collectively exchange prompts, workflows, failures, and learnings.
3. Sharpening priorities. Because AI can make more things possible, it becomes increasingly difficult to distinguish between activity and impact. Leaders who handle this skillfully assist their teams in understanding what is now possible and what should be done.
4. Delegating outcomes instead of tasks. AI-driven work allows space for people to determine the path to completion. The manager sets an objective, creates guardrails, and creates room to explore.
5. Keeping the human experience of the change in view. AI adoption evokes apprehension, exhaustion, and existential reflection. Those managers who accept that, rather than press on regardless, engender in their teams the strength required to remain productive.

How Manager Development Has to Change

Building this capability for scale is what most organizations miss. None of the usual levers (a workshop, a framework, a once-per-year program) were really built for behavioral change in a dynamic environment.

Mento matches coaches who have operating experience — such as former product vice presidents, chief technology officers, and heads of marketing — with an AI layer that the company calls Moments. Moments learns and records key patterns and learnings from coaching so they stay present and the advice can be applied to day-to-day work instead of being left behind after the call.

The specific model matters less than the principle behind it: continuous, contextual learning tailored to where managers work. In Mento’s own data, that strategy lines up with tangible changes in the behaviors most challenged by the stresses of our AI era: more focused long-term planning, better delegation, and the ability to avoid burnout.

Where HR Leaders Go From Here

HR groups feeling the pressure to roll out AI fluency programs quickly might be tempted to frame tool adoption and leadership development as distinct programs. They aren't.

"Don't let the urgency of tool adoption distract from developing the people responsible for making adoption work," Albers said.

Managers serve as the critical bridge between a company’s strategy and its actual operations, with the authority to either enable or stifle a new endeavor’s adoption within their unit. That was true before AI, and it is even more true now. The companies finding success with their AI investments are recognizing that their leadership programs are a performance infrastructure. They are no longer viewing it as just another L&D cost.

The past two years have seen companies asking if their people can handle AI. The question companies should really be asking is if their people managers can manage through AI.

Learn more about Mento.