You stand at the starting line. Your partner is two-legs and a torso split from a surface of plastic and chrome. It’s like the dancing robot from the old Honda commercials, accept it’s bigger, coming right up to your chest. You can hear the powerful hum of a thousand tiny motors under its chassis as it powers up.
We’re not racing AI. We’re tied to it.
You look down and realise you and the robot are tied together at the leg. The starting gun could fire at any moment, and you picture a scene from an old western: a cowboy dragged across the ground by a bolting horse. This could end very badly. What do you do? Don’t hesitate.
The winning strategy for a three-legged race isn’t raw speed. It’s synchronisation and a steady cadence. You move together, find a rhythm, and if you want to reach the finish line, the slower or taller person sets the pace.
That’s where we are with AI. Most organisations are treating it like a thing that should go ever faster. But we are still its slower, taller partner. We set the pace, we see further ahead – for now – and it is on us to lead. It is also on us to stop ourselves and the people around us from falling flat on our faces.
That’s where wellbeing comes into the picture as something that’s essential to an effective partnership with AI. Learning how to move well together is how we’re going to keep up.
This article is all about how AI and wellbeing can work together to enhance each other, helping you get to grips with the intersection of two increasingly integrated topics. And don’t forget to apply the three practical tests at the end of this article to see whether you’re using AI in a way that boosts your wellbeing at work.
Fear undermines both wellbeing and AI implementation
The shadow AI casts over the workplace is one of fear. Leaders fear losing control. People fear falling behind. Everyone is aware that not keeping up may have consequences for their role. That fear is already shaping behaviour in ways that undermine both wellbeing and performance.
In some organisations it shows up as a lack of trust: overly restrictive mandates that don’t grant people enough ownership over the way AI fits into their work. When trust and autonomy drop, wellbeing follows. This makes a second trend worse: people adopting tools in ways that do not genuinely help them, simply to keep up with expectations. Where these trends combine AI costs climb, burnout increases and productive AI implementation slows to a crawl.
Building confidence, not control
The response is not more control, but more confidence. At Impact, we see time and again that trust is built through understanding. Strong partnerships – whether between people, or between people and technology – are formed by listening to what is actually needed. And what organisations, individuals and even the tools themselves are signalling is that work is changing. People need to use AI differently, and in some cases, do fundamentally different work.
Where organisations are making progress, they are doing something subtle but important: they are bringing AI and wellbeing together rather than treating them as separate agendas.
AI can offer timely support for wellbeing
The reality of modern work is that demand continues to increase while human energy does not. AI can help rebalance that. It can step in as a second pair of eyes, support when energy is low or reduce the cognitive load of routine tasks. Used well, it protects time and attention. But the key difference lies in discretion. When people have the autonomy to choose how they use AI – adapting it to their strengths, their context and their judgment – it supports wellbeing rather than undermining it. In successful organisations it’s promoting leadership actions from every level as people take more ownership of their work and when AI should be part of it.
AI when it’s needed, but human where it matters
Knowing when not to use AI is as important as knowing when to use it. Technology is an enabler, but for everything it helps us do, it can also quietly stop us doing something else. There are some things we cannot afford to outsource. Human connection is one of them. Our connection to context, to other people, and to values is critical not only for wellbeing but for the quality of our work. AI might help write an email, but it cannot replace the act of thinking about the thoughts and feelings of person receiving it or the potential consequences of what you communicate.
We can’t let AI get between people, with groups of people only interacting with each other’s AI outputs, because when humans connect to humans they both feel better and understand better. Adaptive organisations are using AI as a bridge to more of that connection, and those that help their people to engage on a human-to-human level are seeing further into today’s challenges.
The secret to high-performing AI is high-performing teams
This has implications for performance. AI might triple its outputs when you double yours, but how long can you sustain that level of work? The marker of effective performance is sustainable effort. If productivity increases at the cost of burnout, it becomes a form of debt that organisations will eventually have to pay back. High-performing AI still depends on high-performing teams for its best outputs, but humans do not scale infinitely.
Keeping humans in the loop is like riding a bike. You can pedal as fast as you like but you have limits, and if you want to go the distance you vary your effort to match the terrain. Think Zone 2 cardio: a steady, moderate pace where you’re working but not straining - effortful, yet comfortable enough to hold a conversation. In knowledge work, this translates to a level of focus where you can stay engaged, briefly step away, collaborate, and smoothly return to the flow of a task. This is where humans perform best. Elite athletes spend 60–80% of their time here, storing up intensity for when it matters.
Designing human-centred AI workflows
We want to stay in that ‘Zone 2’ state so that we have the capacity to easily pick-up what AI still fails at. Thinking reflectively and critically, managing and moving between contexts and weighing their intrinsic stakes are just some of the things we do better than the bots.
Align human strengths with AI’s limitations. We’re full of ideas for example and AI can make it easier to put them into practice, but what about generating them in the first place? Do we generate our best ideas sitting still, typing into a terminal? Probably not.
Research suggests human thinking switches between two modes: a static mode, where we focus deeply on one thing, and an active mode, where we move, scan, and sift for new ideas. The fix is alignment. Use movement for ideation - walking, talking, capturing thoughts with AI voice or transcription. Then switch modes: sit down and focus on one thing at a time. Bring AI back in the loop when you have something concrete to refine or extend.
In practice, it becomes a loop that's better for the way you feel and better for your outputs: generate (AI-supported), execute (human-focused), implement (AI-supported). Living, moving, shifting, executing precisely; this is a style of work that's more effective and more human.
Is your AI use supporting wellbeing? Apply a simple framework to find out
Phew - if you stuck with us from the start that's a lot. It's hard for it not to be when AI is squeezing into every aspect of our work. It means we can't talk about wellbeing separately from or AI, or AI from wellbeing. That's why we've come up with three tests you can apply next time you're considering a use case for an AI and whether it will enhance wellbeing. Remember, if it's not enhancing wellbeing in the long run, then it's probably not enhancing your prospects of doing well in the race for profitable AI implementation either.
To rate whether a particular use case for AI is working for your wellbeing, consider if it’s D – C – P: discrete, connected and paced.
Discrete:
Does this genuinely solve a problem at an acceptable level risk?
Connected:
Does this improve human connection, or free up time for it?
Paced:
Can you sustain this without burning out? Do you have enough left to focus on the other things you need to do well?
If the answer to any of these is no, it is worth pausing.
This isn’t just a race to go faster
We made it! Except this is only the starting line…
The real race starts in workplaces around the world every morning, every new use case, every gain in efficiency, contributes to a much larger shift in how organisations operate. We have been through transformations like this before, and history shows that speed without care comes at a cost – think Industrial Revolution and the wellbeing of miners, chimney sweeps and factory workers. This time, there is an opportunity to do things differently.
The question is not how fast we can move with AI. It is how we choose to move forward together. It’s about how well we understand the work we’re doing.
The future will not be defined by technology alone, but by how well we combine it with human capability. And the organisations that succeed will not simply be the fastest. They will be the ones that know how to go fast, while still going well — because ultimately, that is how they will go further.
In a three-legged race, going faster isn’t what wins. Staying upright is.