AI is exploding inside companies, but transformation is still lagging behind. We covered this in our last article: 89% of global executives see no measurable effect of AI on their company's productivity, and only 12% of employees at organizations that have adopted AI say it has genuinely transformed their work. One data point, however, completely changes how we should read this finding: according to Gallup, it isn't a question of budget or technology. It's the manager's role that tips an organization one way or the other on AI adoption.

AI is booming, transformation is still waiting

A quick reminder before going further. 95% of enterprise AI pilots generated no measurable ROI (MIT Project NANDA, 2025). A survey of nearly 6,000 executives confirms this same productivity gap (NBER, cited by Gallup, 2026). Individual adoption is massive; organizational transformation remains rare. The question is: why the disconnect?

The AI manager factor: the number one driver, ahead of the tech itself

Gallup identified the two main factors that explain frequent AI use within an organization. The first, unsurprisingly, is technical integration into existing tools. The second, far less expected, is active support from the direct manager. It outranks company policy, tool provisioning, or internal communication, even though these are where most of the investment goes (Gallup, 2026).

The effect is massive. When a manager actively supports AI use, their team members are 8.7 times more likely to say AI has transformed their work, and 7.4 times more likely to say it frees up their time (Gallup, 2026).

The problem is that this kind of support remains rare. Fewer than 1 in 3 American employees say they currently benefit from it. In Germany, a separate Gallup study puts the figure at just 21% (Gallup Germany, cited by Gallup, 2026). So this isn't an isolated American problem.

The manager's paradox: expected to drive the transformation while running on empty

This is where it gets complicated. Managers are being asked to drive AI adoption within their teams at the exact moment their own engagement is collapsing. Manager engagement has dropped 9 points since 2022, from 31% to 22% (Gallup, 2026). Caught between leadership expecting results and teams needing reassurance, the middle manager is absorbing a rapidly expanding workload, often without the means to match it.

And yet, managers aren't the ones dragging their feet on AI. In France, 55% of managers now use it at least once a week, a 15-point increase in one year (Apec, 2026). Managers are adopting the tool for themselves, often faster than the rest of the organization. The real gap lies elsewhere: knowing how to use AI for yourself and knowing how to bring an entire team along are two completely different skills. Without structured support, a manager can be perfectly comfortable with ChatGPT for their own use while lacking the space and the setup needed to help their team.

What a manager who "actively supports" AI actually does

This question deserves more than a paragraph. "Actively supporting" AI adoption isn't sending a link to a tutorial or activating a ChatGPT license for the whole team. It's a posture, a set of rituals, a concrete way of guiding change day to day. That's exactly what Pierre Baron, CPO at Yago, had to build from scratch in his own organization, including his wins, his mistakes, and what he'd do differently.

The challenge is no longer technology, it's management

AI is here. Your employees are already using it, and so are your managers. What's missing isn't another tool. It's a clear view of what's really happening within each team, so managers can play this role without drowning in dashboards or carrying this transformation alone.

🎥 Pierre Baron, Chief People Officer at Yago, explains concretely how he equipped his managers to face AI: what worked, what he'd do differently, and his advice for organizations just starting out. Watch the full replay of the conversation with Margot Wuillaume (eBloom).