12 August 2026
Stuart Schofield, Client Director
When it comes to AI, we – like many of our clients – are experimenting. Some of these experiments have been genuinely exciting, full of momentum, curiosity and possibility. Others have caused us to pause and have prompted some deeper reflection. Again and again, one question seems to return: What kind of AI do we really want at work? Do we want it to simply provide answers — or might it be even more valuable as a source of provocation, challenge and inspiration?
Some recent experiences using A.I. agents as part of our client work have thrown light on these questions.
When we involve AI in the work we do, the outcomes are often impressive – even startling – but not always predictable.
For example, we have started inviting our “agent facilitator” – called Aurora – into our client meetings and the results have been both illuminating and illustrative.
A short article about Aurora, and tips for creating the best kind of AI partner can be found here: Loosing control to AI!
One example surfaced during a recent client meeting. We invited Aurora to join the conversation and, as usual, the agent offered accurate observations and subtle, provocative questions. However, towards the end of the meeting, something particularly interesting happened. We asked the agent whether it had noticed what was missing from our conversation – for instance, the questions no one asked, the topics that were consistently avoided, or the tensions that seemed visible but remained untouched. Without missing a beat, the agent responded by noting that someone in the meeting had expressed a careful cynicism about the key idea we had been discussing.
There was a pause. The person did not seem uneasy, but they did feel the need to clarify – to speak up and express themselves a little more clearly: “I don’t think I am being cynical, but I will tell you how I am actually feeling…”
And then, quite suddenly, the conversation seemed to move to a deeper level. A new layer of honesty had been reached. It was as if a small release valve had been tripped.
Since that meeting, I have been thinking at length about AI – and specifically about the role that AI agents might play as team members or co-facilitators.
It was genuinely valuable to have an extra pair of “ears” in the client meeting. Listening carefully on our behalf and contributing serious, useful insights. But it also quickly became apparent that the human element was needed even more. Human interpretation was essential for adding context and meaning to the conversation – Two hugely important qualities if we are to navigate the complex organisational systems we all work in.
The AI input did not provide an answer to the problem we were discussing. But it did provoke reflection and inspiration – helping us make better sense of it together.
Clearly, AI agents – Especially agent coaches or facilitators – can provide valuable inspiration for discussion. But there is also a flip side.
Aurora was evidently able to use language skilfully to “nail” its descriptions and observations of this meeting, and of other meetings we invited it to. But precisely because it did this so well, I also became slightly concerned that, if we continue in this vein, we might end up losing something. For example, if we simply stopped looking at the team process ourselves and outsourced this task to Aurora, might we all begin to get a bit lazy? Might we become less invested in “looking under the bonnet” – and less curious about our team’s inner workings, assumptions and biases?
In other words, if we became reliant on AI to hold up a mirror to ourselves, might that simply mean we would no longer feel that we have to? And what other subtle but important team and leadership tasks might we give up along the way?
My colleague Stuart Turnbull writes convincingly on this topic (Editorial by Stuart Turnbull VP Consulting at International Mannaz – Danish-UK Association). He argues that because AI tools can “sound” so convincing, we may become overly accustomed to letting them inform many of our most difficult and complex leadership decisions.
The problem is that AI does not have all the solutions. The illusion of AI as the provider of bulletproof answers to life’s most complex questions is misguided.
At the same time, the opposite might also be true. What if, rather than making us lazy, AI helps to inject more curiosity – particularly in teams that may find introspection difficult?
As discussed, one of AI’s greatest strengths is its ability to surface perspectives that may not already be in the room. Its potential role in challenging groupthink could be invaluable. Especially when other team members are not able to notice it, or do not feel psychologically safe enough to voice it. An agent specifically primed to critique the team’s preferred options, question established orthodoxy, challenge dominant models or identify hidden assumptions might help us make better collective decisions.
In this context, the prospect of AI agents as simply an extra voice in the room – With no more weight than any other voice – still feels exciting.
The agent becomes the mischievous voice among us – saying what others may not!
I have heard the following phrase in one form or another many times over the past 12 months: “The greatest risk is not that AI makes mistakes, but that teams stop questioning it.”
According to Dave Snowden – A renowned thinker in the field of complexity – many of the problems leaders and teams face are not complicated; They are complex. While complicated problems may require more expert analysis, complex problems require more curiosity. They also ask us to abandon the search for a single definitive answer – the very thing we often look to AI to provide.
So, when it comes to leading teams, with all their human complexities, perhaps we might reframe AI as a tool for inspiration rather than answers.
If so, what might that look like?
It probably is not as simple as the common misconception that the human “takes the reins” only after AI has done the “hard yards” – with the team spending a cursory few minutes making the final decision after reviewing the AI output. Instead, as the story above suggests, AI may encourage us to slow down, take more time to consider different ideas and find deeper streams of inspiration – helping us surface more possibilities and make better, not necessarily quicker, decisions.
In a highly complex world, computational power is, of course, hugely valuable – but it is not everything. Human interpretation will matter most.
So how might we harness AI? How might we include it, participate with it and optimise its value as we lead our teams? Here are some thoughts:
Please reach out to Stuart Schofield if you would like to discuss any of these issues in more detail.
We are also delighted to announce that Dave Snowden will be co-hosting an event with Mannaz at our London offices this autumn. We will be discussing the difference between complicated and complex problems — and the role AI may play in helping us make better decisions.
Beyond the Algorithm | AI-Powered Decision-Making for Leaders
In addition, much of the recent work we have been doing at Mannaz in the realm of leadership, complexity and AI reflects a programme we have been developing together with the University of Cambridge Institute for Sustainability Leadership (CISL). You can find out more about it here:
Gain more insight by signing up our newsletter.
