Mervyn Dinnen
Writer / Speaker / Podcast Host / Analyst / Top 100 HR Tech Influencer / co-author of the books Digital Talent & Exceptional Talent
The questions most organisations are asking about AI is the wrong one:
What is AI doing to our culture?
How do we manage the impact?
How do we bring our people with us through the change?
These are reasonable questions. But, according to Nick Holmes – a year into a PhD on organisational culture and one of the more original thinkers I’ve spoken to on this subject on the latest episode of the HR Means Business podcast – they all start from the wrong premise.
AI is a tool. And when you put a powerful tool into an environment, the environment determines what happens next.
The tool doesn’t.
The better question is not what AI is doing to your culture. It’s whether your culture is designed to make use of it at all.
The Familiarity Trap
Nick’s PhD has surfaced a finding that should give every HR leader a pause. There are 164 different definitions of organisational culture and 70 different measurement tools – and none of them have reached consensus. Culture is, as he puts it, treated atmospherically: like the weather, ever-changing, hard to pin down, something everyone references and almost nobody can truly explain.
This is not an academic problem. It is a practical one. Leaders who cannot explain how culture actually works cannot design it, cannot measure whether it’s changing, and cannot create the conditions AI requires to deliver value.
The zip analogy makes it concrete. Everyone can use a zip. Almost nobody can explain the mechanism – how the teeth engage, how the lock works. The moment you ask, confidence collapses. AI is amplifying exactly this effect inside organisations. People feel more familiar with subjects after a ChatGPT query. They don’t understand them more deeply. And leaders are making AI strategy decisions – agreeing to roadmaps, signing off on investments – based on familiarity rather than a genuine understanding.
The Engagement Coincidence
Gallup has tracked employee engagement for nearly 26 years. When they started, it sat at 23%. Today it remains at 23%. In that same window, and specifically over the last five years, AI adoption has accelerated while engagement has declined.
Nick was careful not to claim causation. But he doesn’t dismiss the coincidence either. His explanation is structural: organisations are no longer running large, discrete redundancy exercises. They are cutting continuously – a few roles here, a few there, constantly.
The psychological research is clear that it takes around six months to recover from the impact of a layoff round. When cuts are continuous, organisations are permanently in psychological deficit. Trust never fully rebuilds. And employees completing engagement surveys while that cloud hangs over them will answer negatively, even when the work itself is unchanged.
The Sabotage Nobody is Talking About
I think there’s one topic from our conversation that deserves to be taken seriously at leadership level – according to the Workplace Intelligence Report 2026, 29% of employees admit to actively undermining their employer’s AI strategy. Among Gen Z, that figure rises to 44%.
This isn’t resistance. It’s sabotage. And the generation most assumed to be AI-native is the most likely to engage in it – precisely because they are the most aware of what it can do and the most anxious about what it might mean for their futures. A survey of 17 and 18 year olds found that 70% would prefer to have grown up without social media and AI entirely.
The implications are significant. No AI strategy survives 44% of its youngest employees actively working against it. The response cannot be enforcement. It has to be trust – and trust is a cultural outcome, not a communications campaign.
The Cognitive Load Problem
The promise of AI has always been that it would free people to do more human, creative, and meaningful work. The data suggests something different is happening. 77% of employees say AI has increased their workload. Frequent AI users report 45% higher levels of burnout.
The reason, Nick argues, is that organisations are using AI to do things faster without stopping doing the things underneath. Work accelerates. More gets added. Zombie processes survive because nobody asked what should stop. The organisations genuinely extracting value from AI ask a different question: not what can we do faster, but what can we stop doing entirely?
What Good Looks Like
The organisations that Nick sees getting this right share one characteristic – they don’t lead with efficiency. They lead with purpose.
Their message to employees is not “AI will help us work faster” – nobody gets out of bed for efficiency. It is: “How do we use these tools to enhance what you already do brilliantly, and change what our organisation is capable of?”
That distinction may sound small. But its consequences are not.
Action Points for HR Professionals
Based on my conversation with Nick, here are five practical starting points:
Audit cultural readiness before AI readiness. Before your next AI investment, assess the environment it will land in. Do your people have psychological safety? Is trust between managers and teams strong? Is change communicated well or poorly? The answers determine AI outcomes more than the technology choice does.
Apply the McKinsey ratio. For every pound invested in AI tools, plan to invest five in your people – in upskilling, change readiness, adoption support and manager capability. Most organisations are spending the inverse of this.
Take the sabotage data seriously. If a significant proportion of your workforce – and particularly your younger employees – are actively working against your AI strategy, a communications push will not fix it. Understand the specific fears driving it and address them directly.
Ask what stops, not just what starts. Before implementing AI in any workflow, ask explicitly: what existing process or task will this replace entirely? If the answer is nothing, you are adding cognitive load, not reducing it.
Train managers first. Managers are the lynchpin between AI strategy and employee experience. If they are not equipped to use these tools themselves, cannot answer their teams’ questions about them, and are not modelling their use confidently, adoption will stall at every level below them.
You can check out my full conversation with Nick Holmes here – https://www.hrhappyhour.net/episodes/how-culture-can-make-or-break-ai-adoption/ – or through the image below