Companies everywhere are trying to figure out how to get their employees to use AI. They are investing in new tools, developing AI strategies, running training programs and encouraging people to experiment. According to the BCG’s AI Radar Survey, leaders are expected to invest approximately 1.7 percent of their annual revenue in AI in 2026.
And yet, AI adoption remains a stubbornly difficult challenge.
A recent Harvard Business Review article by Gleb Tsipursky offers an interesting explanation. He argues that senior leaders often assume AI adoption requires three things; a clear strategy, the right technology and workforce training. All three matter, but they are not enough.
Somewhere between the strategy announced by senior leaders and the everyday behaviour of employees, middle managers have to translate AI ambitions into new ways of working.
The Missing Layer of AI Adoption
Middle managers occupy an interesting position in any transformation. They are close enough to senior leadership to understand what the organization is trying to achieve, and also close enough to employees to shape how those ambitions are actually implemented.
This makes them particularly important for AI adoption. Employees might hear executives promise that AI will eliminate boring work and improve productivity, but they will look to their managers to understand what those promises actually mean.
-
Will workloads change?
-
Will performance expectations change?
-
What happens if AI produces something inaccurate?
-
Are people genuinely allowed to experiment and make mistakes?
In other words, senior leaders can make AI a priority and technology teams can make AI available, but middle managers are the ones who make AI real.
This is why I think we need to stop thinking about middle managers as simply the implementation layer of an organization. During transformation, they are translators.
They take something abstract and make it practical.
Translation Requires Buy-In
Middle managers cannot effectively translate a transformation they don’t buy into. And buy-in is more complicated than simply asking whether somebody thinks AI is a good idea. In my work, I think about buy-in along two dimensions; belief and participation. Do people believe in the change you are trying to make? And are they actively participating in making it happen?
If you have one and not the other, you don’t have buy-in.
A manager might believe AI has enormous potential but disagree with how the organization is implementing it. Another might enthusiastically experiment with AI but lack the confidence or authority to change how their team works. And another might comply with an executive mandate while privately believing the whole initiative is misguided.
This is why treating middle managers as one homogeneous group doesn’t work. Tsipursky’s research similarly identifies five different middle-manager profiles, each responding differently to risk, evidence, incentives and support. The leadership challenge is therefore not simply getting managers to use AI. It is understanding where their buy-in currently sits and what they need in order to move.
Conclusion
Middle managers have always mattered during organizational transformation. The speed of change driven by AI has brought their role more strongly into the forefront. Senior executives who are looking to drive AI adoption should play close attention to whether they have buy-in from their middle managers.
If you have a transformation initiative you are working on, we can help you to create the social capital you need to get your ideas adopted.
Join my masterclass to learn and apply practical tools to create the gravity that attracts people to willingly support and actively participate in change.

