You’ve spent the money. You’ve equipped your people with all the latest tools (you might even have built some of your own). You’ve mandated the training, and the productivity / performance / innovation benefits speak for themselves. Yet scale sounds elusive.
Sounds familiar? You’re not alone. McKinsey data shows us that 65% of organisations are using generative AI in at least one function, yet only 11% have scaled it across their business1. That’s because AI transformation asks more of us than any technology rollout in history. It has shifted the lines between human judgement and machine assistance and unsettles how people define their value. It even questions the value exchange between businesses.
Working within Havas’ AI accelerator for the past two years, I’ve had a close view how this played out for us in over 100 markets, alongside watching clients on their journey. What I’ve come to learn is that’s less about the tech, and more about the people. It’s how they lean into… or hesitate around… the tech. It’s the tiny pause before using a tool in front of others. It’s the cautious glance as to whether this is the ‘done’ thing.
People always read the room first – it’s just how we work. You notice who’s giving it a go, who’s quietly capable, and who’s praying they won’t be asked. This isn’t new, but AI has made it more obvious. Some jumped in because they enjoy the challenge, but most sat in a careful middle ground, trying to figure out what all of this means for their judgement, pace, relevance and future.
Here’s where leaders can play a key role. To remove any potential anxiety or fear around AI in the workplace, they must lead from the front – encouraging and motivating people to see the value and benefit of using new technology. This this isn’t just a theory; it’s a well-researched method for success. Studies from MIT Technology Review Insights 2 showed that employees are far more willing to experiment with new tools when they see leaders modelling curiosity and openness, rather than uncertainty or reluctance. Fundamentally it’s human nature – people like the idea that we’re all figuring this out together – removing any fear of failure or expectation.
But crucially, teams don’t want or need hype. They need a steady tone and a bit of room to work things out without feeling watched. A calm, honest, ‘let’s see where this actually helps’ travels fast because it feels believable.
Momentum comes from the practical rather than the polished. The teams that adopted AI the quickest were the ones that made it normal to talk about, not just easy to use. A rough walkthrough lands better than a training deck because it feels real, not theoretical. And once somebody gets a small win – shaving time off a task, sharpening a thought, speeding up a draft – the whole thing starts to roll on its own.
When I speak to clients, the ones making the fastest progress are the ones that acknowledge the human challenges as well as the machine ones. They are the ones treating AI adoption as a behavioural shift as much of a technical one – and that small but significant mindset change shows up everywhere.
Confidence spreads when people feel safe enough to show their working. Ideas surface when people aren’t bracing for judgement. The fastest progress usually comes when leaders normalise the whole cycle; what worked, what didn’t, what we learned, and what we’ll try next. Keep it lightweight, keep it visible, and let the culture you create do the heavy lifting.
The tech will keep shifting, but progress still depends on whether people feel safe enough to try things in front of each other. The leaders seeing the most traction treat this like any other change programme; they anchor AI to real work, redefine what ‘done’ means and reward people who share what works early. Without structure, adoption stays performative. [1] [2]
