For as long as I've been in business, we've designed our organisations, processes, capacity and scale around two things: people and tools - IT, automation and equipment.
Tools gave us scale and did what they were designed to do, reliably and predictably. You could slot them into or across processes or tasks and trust them to behave the same way every time. People brought something tools never could: judgement, empathy and the ability to read a situation, weigh competing priorities and make a call.
For decades, that's been the architecture of organisations. A deterministic piece, and a human piece. Everything that’s been built on top of that, from organisational structures to governance and operating procedures, has been designed around that split.
AI doesn't sit comfortably in either box. It has the scale and speed of machinery and tech, but it can also consider context and make decisions in ways that aren't deterministic. That's different from anything we've previously introduced into a business process. It doesn't behave like the tools we've always known, but neither is it a person we physically interact with, coach, observe, read body language or instinctively trust.
It's a completely new capability sitting inside our organisations, and it's why leaders find themselves slightly off balance. Our instincts keep reaching for familiar patterns and playbooks, but they don't quite fit. We can't govern AI like a new machine or deterministic piece of tech, and we can't manage it like a member of the team.
We've lived through these seismic shifts before, with advances like electricity, computers, the internet and mobile technology changing the way businesses operate. The time feels different because AI has become widely available and accessible almost overnight. Leaders are trying to understand it, make decisions about it and lead other people through it while they're still learning themselves.
The epiphany, and what it actually is
So how do you get comfortable with something that doesn't fit the models you've relied on for years? Here's what I've learned: you don't read your way there. You have to go and experience it in a way that’s relevant for you.
One of the most useful things I do, and recommend to any leader wrestling with this, is to deliberately envision the utopia version of what AI and agents can do inside a business process.
Not because it's realistic. It isn't, at least not yet. There's still work to do to build the right foundations: data, guardrails, security, observability, process redesign and change management. And the technology itself is still evolving.
But looking at the aspirational version does something important. It breaks the mould of how you think about your business and the way you’ve traditionally thought about how you structure teams, workflows, interactions and processes. It gets you up off the dance floor, where you're too close to see anything but the next step, and onto the balcony, where you can see the whole shape of what's possible. Getting onto the balcony doesn't happen by accident though.
You need enough curiosity to explore what AI is capable of, and enough understanding to start seeing the opportunities. Sometimes that comes from your own experimentation. Sometimes it comes from seeing how others are applying it. Either way, you need some idea of what you're looking at before you can imagine what's possible for your business.
That's probably the closest thing I've had to an epiphany. Not a flash of insight that suddenly resolved everything but a change in vantage point that helped me ask better questions.
I'll give you a simple example. You ask AI to analyse last quarter's sales and compare them with the previous year. The report comes back looking polished and convincing, but you can tell something doesn't quite add up. You investigate and discover the data isn't grounded properly. You fix that, run it again and it's still not right because it needs more business context.
You start with the balcony view of what's possible. Then you experience first-hand why getting AI to produce reliable outcomes is harder than you first think. That's when things like guardrails, semantic models and governance stop being technical concepts and start making practical sense.
There’s another aspect of this I think is important to be open about. As leaders, we still don't know enough about AI, or its long-term impact on the bottom line, to stop being sceptical. We should be sceptical. There are real questions about which tools are best, where our data ends up, how we manage cost and how we deal with things like hallucinations.
I don’t feel like that’s a failure of nerve. It's a reasonable response to something we're still learning how to apply as a transformation tool even while it's already proving itself in specific, well-chosen use cases.
What I’ve found is that you have to hold both views at the same time. The optimism and excitement that comes from standing on the balcony and seeing what's possible, and the healthy scepticism that comes from knowing how much still needs to be learnt.
Balancing those two ideas doesn't come naturally. We're used to being the people with the answers, yet we're being asked to spend longer with open questions than feels natural and to bring our teams along with us before the answers are clear.
Once you've been through that journey yourself, you begin to recognise it in other people. You can tell who's taken the time to get beneath the surface of AI, who's genuinely rethinking the way work could be done, and who's still looking at AI through the lens of existing ways of working.
As a leader, that changes the conversations you have. You're not just looking for people using AI. You're looking for people whose thinking has changed because they've taken the time to understand it. They're the people who begin imagining different ways of working, and they're often the ones who help the rest of the organisation move forward.
The identity underneath
There's a behaviour I see in myself, and other leaders I know, that actively works against us here. When something new and uncomfortable comes along, our instinct is to make sense of it using the experience that's served us well for years. We look for familiar patterns and ask ourselves, "How would I normally solve this?"
It feels sensible and it feels like leadership. But it's also how we can miss the opportunity AI is putting in front of us. The moment we filter something genuinely new through the lens of ‘how we've always worked’, we've already started shrinking it back into something familiar that doesn’t really change anything.
Part of why that instinct is so strong is that AI challenges one of the things leaders have always relied on most: our ability to govern.
With people, we know how to do this. We have policies, processes and controls. More importantly, we can sit across the table from someone, listen to how they think through a problem, evaluate their judgement and build confidence in it over time.
With technology, governance has always been different but equally familiar. It was deterministic. You tested it, maintained it and expected it to behave consistently.
AI gives us neither of those comforts. It can demonstrate judgement like a person, but we can't sit across the table from it and understand how it reached a conclusion. The observation layer we've relied on for years simply isn't there in the same way. That's genuinely unsettling, and I don't think it's irrational to feel that.
Underneath all this sits something even more personal. Think about what makes a great salesperson, consultant or customer service professional genuinely valuable. It isn't just knowledge of your business process. It's judgement and EQ developed over years, knowing when to challenge, when to adapt and when to make a call that others trust. That experience and expertise have often been what people have built their careers on.
AI's real opportunity is that it can begin capturing and scaling some of that expertise. It can make the judgement of your best people available more consistently across the organisation, rather than relying on where those individuals happen to be.
That's an extraordinary opportunity for a business. It's also, honestly, a little confronting if that expertise has been your personal edge for twenty years.
I don't believe that makes our experience less valuable. If anything, it makes it more important. The difference is that where we create value starts to shift. It becomes less about personally holding the answers, and more about deciding which judgement should be embedded, where human judgement still matters most, and how the two work together.
What I'm taking from all of this
For me, none of this changes the fundamentals of how we should make decisions. AI still has to solve real business problems. It still needs a business case and still needs to create P&L value. The discipline we've always applied to investing dollars and effort into anything doesn't disappear because something is labelled AI.
I’m also recognising it needs courage. The courage I mean is being genuinely open to rethinking an entire linear process, and accepting that it might eventually change the shape of your org chart. That brings uncertainty, especially when not everyone around you is doing it yet. But there are enough proven use cases now to start somewhere real, demonstrate value and build the confidence to go further.
The clearest example for me is our own business. Look at how we run a sales cycle and develop and deliver solutions for customers today. What we do now is already very different to what we were doing six months ago. Our output is stronger, the quality of our bids and submissions has improved, and we're tying that directly to win rate, not vaguely, but as a metric we actively track.
Better solutioning means better value for the customer and a better win rate for us, which means we can grow revenue without scaling sales costs in the same way. We're still early in that data, but the trend is real and we watch it constantly. This isn't a set-and-forget exercise. It's about continually testing, learning and understanding where AI is genuinely creating P&L value, not just the illusion of productivity.
We've adapted to major shifts before, and we'll adapt to this one as well. In most careers there are only a handful of moments where technology genuinely changes how organisations operate. I believe this is one of them.
That's why my own approach hasn’t been to wait until every answer is clear. It's to lean in, stay curious and keep learning. The difference this time is that we're learning while we're leading. That does feel uncomfortable, but I suspect it's exactly what this next chapter demands of us.
Sven Martin | Executive Director | Business Applications
Sven Martin has spent more than 16 years at Fusion5 in a range of senior leadership roles. His experience across business applications, client engagement and organisational leadership gives him a broad perspective on how systems, data and strategy come together to drive meaningful business change.
This article shares one executive's reflections on leading through AI.
If you're exploring what it takes to move AI from experimentation into production, these companion resources take a deeper look at the organisational capabilities, governance and operating models required.
