Skip to content
ADVERTISEMENT
Recruitment & HR

Middle managers hold the key to AI transformation 

Alice Desbouvrie discusses the role HR and middle managers can play in turning AI investment into meaningful workplace adoption.

Opinion
4 min read
Share
Lumapps alice desbouvrie
ADVERTISEMENT

Most organisations are further along with artificial intelligence (AI) than they give themselves credit for. Investment is up, with more than half of UK firms (54%) actively using AI,  and in many cases, the tools are already in people’s hands. The harder question, and the one HR is increasingly being asked to answer, is how to make those returns match the ambition.

The answer, more often than not, sits in the layer between strategy and delivery.

Research published in the Harvard Business Review captures something worth taking seriously. Senior executives and middle managers have different understandings of what AI demands for their business. Executives are, rightly, focused on the strategic opportunity, whereas managers are often navigating more of the operational reality, which appears messier and far less forgiving of a poorly supported rollout. Closing that gap is one of the most concrete things HR can do to improve AI outcomes, and it’s more tractable than it might look.

What the data is actually telling us

It’s easy to read slow adoption as disengagement. However, the data suggests something different: BCG’s AI at Work 2025 report found that managers use of AI has risen to 78% in 2025, up from 46% in 2018, while frontline adoption has stalled at 51% and hasn’t meaningfully moved in two years. The gap isn’t about willingness. It’s about who has been given the context, the tools and the time to actually integrate AI into how they work.

ADVERTISEMENT

For HR, that’s actually an encouraging insight as it points towards the problem being more structural than cultural, and structural problems can be fixed. Unlike mindset or attitude, systems, processes and support structures can be redesigned, retested and improved over time.

The view from the middle

Middle managers are rarely the problem – but they’re frequently treated like they are.

ADVERTISEMENT

Managers are often sceptical of poorly supported rollouts for good reason, as they’re the ones carrying the consequences. They know which team members are struggling and might not say so. They understand the difference between surface adoption, using a tool because it’s been mandated and integration that actually changes how work gets done. Their pushback, where it exists, is usually a signal worth reading rather than some form of resistance to overcome.

The Wharton research is explicit on this point: It’s organisational readiness in the form of leadership alignment, workforce skills, governance, and change management. Managers sit at the intersection of all of those. The organisations getting the most from AI are the ones treating that layer as a strategic asset rather than a delivery mechanism.

ADVERTISEMENT

Three things HR can do now

ADVERTISEMENT

To help close this gap, it’s about changing the conditions in which managers operate.

Most adoption dashboards track completion rates and headcount usage. Neither tells you whether managers actually have the confidence or context to lead AI-enabled teams. A more useful measure shifts the question from ‘Did your team complete the training?’ to ‘Has your team changed how they work?’ Building that into line manager objectives and giving managers visibility of where their teams actually stand helps to create a different kind of accountability, one that might currently not exist in many organisations.

The translation problem matters just as much. Managers need to understand not just that AI is a priority, but what it means for their team’s work in the next quarter. That connection between strategic intent and day-to-day application needs to be designed, and HR is well-placed to own it.

The third shift is about how managers learn. Often, most get the same generic awareness content as everyone else, then are expected to share it with their teams with little else to go on. Bringing managers together across departments to work through real challenges, rather than hypothetical ones, tends to stick in a way that a module simply doesn’t. The conversations that happen in those sessions often matter as much as the content itself.

ADVERTISEMENT

The opportunity is real

Skills England has set a target of upskilling 7.5 million UK workers by 2030, and the Wharton study found four in five executives expect AI investments to pay off within two to three years. Both of those timelines are achievable, but only if the alignment between executive ambition and management reality is treated as something that needs active design, not only passive assumption.

The good news is that the conditions for this are better than they’ve ever been. Executives want results. Managers want support. Workforces want to develop. HR sits at the centre of all three.  The organisations that treat middle management as a lever – not a lag – are the ones that will turn AI investment into real, measurable impact. HR is uniquely placed to make that happen.

Alice Desbouvrie is chief people officer at LumApps

Related