Why The Businesses That Succeed With AI Aren't The Ones That Saw It Coming
The trait that separates businesses thriving with AI from those still stalling isn't foresight - it's curiosity. Here's what nine years of the conversation taught me.
In 2017, I gave a talk on artificial intelligence at a financial services conference. I played a video by futurist Gerd Leonhard about exponential change - self-driving cars, computers that can learn and think, automation reshaping how we work [1]. Then I laid out what I thought was a fairly measured case: AI would start reshaping professional services within five years, and within ten, the changes would be sweeping.
The room was polite. But I could feel it. Most of the audience thought I was describing science fiction. The technology felt intangible - not here yet, not relevant to their Wednesday morning. I left feeling like I'd misjudged the audience. Maybe I had. But I hadn't misjudged the timeline.
What happens when smart people meet discontinuous change?
Two years later, at the 2019 New Zealand Financial Services conference, I watched a financial adviser stand up and deliver an impassioned rebuttal to a guest speaker. AI was not coming for their industry. It was never going to have a material impact on how they served their clients. He was articulate, experienced, and completely confident.
And here's the thing - he wasn't wrong to be sceptical. He was a smart, experienced professional making a rational judgement based on everything he could see in front of him. In 2019, there was no ChatGPT. There was no mainstream generative AI. The tools that would eventually validate everything the speakers had been saying were still three years away from public consciousness.
Everett Rogers, whose research on how innovations spread has been cited across disciplines since the 1960s, made an important distinction between continuous and discontinuous innovation. Continuous innovation is a refinement of something you already know - a faster computer, a better spreadsheet. Your existing mental models still work. Discontinuous innovation breaks those models. It requires entirely new ways of thinking about what's possible [2].
AI in professional services is a discontinuous change. And discontinuous change is precisely the kind that experienced, successful people are worst at anticipating - because their experience tells them the current model works. Which it does. Right up until it doesn't.
“The businesses I’ve seen move most effectively with AI aren’t the ones that had a head start. They’re the ones where the leadership team brought three specific qualities to the table: humility, openness, and curiosity.
Humility to acknowledge that this technology might change the fundamentals of how their firm operates, even if they can’t yet see exactly how. Openness to sit with uncomfortable questions about their business model without immediately reaching for reassurance. And curiosity to explore, experiment, and learn - rather than waiting for someone to hand them a finished answer.”
Why didn't people listen earlier?
Between 2017 and 2022, I talked about AI a lot. Probably too much. At one point, senior colleagues asked me to stop. For years, the conversation felt like pushing against a closed door, because everything I was describing was still, in most practical senses, theory. It wasn't in anyone's daily workflow. It wasn't affecting anyone's revenue. The evidence was directional, not experiential.
And I'll be honest - even I wasn't entirely sure what we should be doing to prepare. I could see how AI would reshape service delivery, shift competitive dynamics, change what firms could do with their data, and raise the bar on output quality. But I was by no means crystal clear on the specifics. The technology wasn't available yet, let alone mainstream. What I was really advocating for wasn't adoption - it was attention. Budget some time. Allocate some resource to learning, investigating, and preparing. That's a much smaller ask than "transform your business" - and yet even that felt like too much for most organisations at the time.
This is worth understanding, not as a story about who was right, but because the pattern repeats. Organisations don't resist new technology because they're incurious or backward. They resist it because investing time and resource in something unproven requires a leap of faith before the evidence is personally visible. Not full-scale adoption - just learning, investigating, preparing. But even that allocation feels hard to justify when there's nothing tangible to point to. Professional services leaders are trained to make evidence-based decisions. When the evidence isn't there yet - when it's research papers and conference talks rather than tools on your desktop - rational scepticism is the default response.
Then ChatGPT went mainstream in late 2022, and the dynamic inverted overnight. Suddenly everyone wanted to know more. They wanted policy, process, training. The same conversations that had been met with resistance for years were now urgent priorities. But even then, the responses split. Some leaders leaned in with curiosity. Others responded with fear, or with a defensiveness that looked a lot like the scepticism from before, just dressed differently.
Does it matter if you were early or late to AI?
Here's the honest answer: not really. Being early didn't give me any particular advantage in implementation. What it gave me was time - time to sit with the possibilities, to think about what the opportunities and risks might look like, to consider where AI could create genuine commercial value. None of that thinking produced a finished plan. But it built a mindset. And that mindset means I'm not scrambling now to learn or believe - I've had years of quiet preparation for a moment that's finally here.
It also gave me time to watch how organisations respond to change they didn't choose - and that taught me something more useful than any technical knowledge.
The businesses I've seen move most effectively with AI aren't the ones that had a head start. They're the ones where the leadership team brought three specific qualities to the table: humility, openness, and curiosity.
Humility to acknowledge that this technology might change the fundamentals of how their firm operates, even if they can't yet see exactly how. Openness to sit with uncomfortable questions about their business model without immediately reaching for reassurance. And curiosity to explore, experiment, and learn - rather than waiting for someone to hand them a finished answer.
Jim Collins identified this pattern in his research for Good to Great. His team studied 1,435 companies over five years and found that the leaders who drove lasting, significant change shared a consistent trait profile: personal humility combined with fierce professional resolve. Collins called them Level 5 leaders. They weren't the loudest or the most charismatic. They were the ones willing to confront uncomfortable evidence about their organisation's position and act on it, without ego getting in the way [3].
That research was published in 2001, well before anyone was talking about AI in the boardroom. But the finding translates directly to what I'm seeing now. The firms that are making real progress with AI aren't led by people who claim to understand the technology. They're led by people who are comfortable saying "I don't fully understand this yet, but I can see it matters, and I want to work out what it means for us."
What separates the businesses that move from the ones that stall?
The difference comes down to how leadership responds, not when they respond. Timing matters far less than posture. And the posture that works is curiosity - a genuine willingness to explore what AI means for the firm's specific situation, rather than defaulting to either panic or denial.
It's tempting to draw a line between "early adopters" and "laggards" - the language of technology adoption curves. But that framing misses what's actually happening on the ground with AI in professional services.
The firms that stall tend to share a pattern. Leadership treats AI as an IT project, or delegates it to someone junior, or commissions a policy document and considers the job done. The underlying assumption is that AI is something to be managed and contained rather than understood and explored.
The firms that move treat AI as a strategic question. They put it on the leadership agenda. They invest time - their own time, not just their team's time - in understanding what it means for their specific services, their clients, and their competitive position. They ask hard questions about where their firm is genuinely exposed and where the real opportunities sit.
None of that requires technical expertise. It requires the willingness to look honestly at where you stand and make decisions based on what you find - even when the answers are uncomfortable. I've seen this first-hand: a firm's readiness to adopt AI correlates far more strongly with the curiosity of its leadership team than with its technical maturity or the size of its IT budget.
What does this mean for your business?
That financial adviser at the 2019 conference wasn't wrong about everything. He was right that AI wasn't going to replace the relationship-driven, judgement-heavy work that good advisers do. What he missed was that AI would change the economics around that work - the research, the compliance, the administration, the client communication - in ways that would reshape what clients expect and what competitors can offer.
If you're reading this and thinking "I probably should have started paying attention to this earlier" - welcome to the club. Almost everyone feels that way. But that thought is less important than what you do next. The window for curiosity-driven exploration hasn't closed. But it is narrowing, because the firms in your sector that are exploring now are building advantages that compound.
The question isn't whether you saw this coming. It's whether you're willing to look clearly at where your firm stands today and make honest decisions about what to do about it. That's a leadership quality, not a technology skill. And it's available to anyone who chooses it.
If you're ready for that conversation, Binary Refinery's AI Strategy & Capability Baseline Workshop is designed exactly for this moment - a structured, facilitated session that gives leadership teams a clear picture of where they stand and what to focus on first.
Sources:
Leonhard, G. (2017). "Digital Transformation: Are You Ready for Exponential Change?" - Watch below. This is the video I showed at that 2017 conference talk. It was made nine years ago. Watch it and tell me how much of it has come true. That's the thing about exponential change - it feels like science fiction right up until it's your Tuesday morning.
Rogers, E.M. (2003). Diffusion of Innovations, 5th Edition. Free Press. Rogers' framework distinguishes between continuous innovations (refinements of existing products and practices) and discontinuous innovations that require adopters to develop entirely new behaviours and mental models.
Collins, J. (2001). Good to Great: Why Some Companies Make the Leap... and Others Don't. HarperBusiness. Collins' research team studied 1,435 companies and identified Level 5 leadership - characterised by personal humility combined with professional will - as a consistent factor in companies that made sustained transformations. Originally published as "Level 5 Leadership: The Triumph of Humility and Fierce Resolve" in Harvard Business Review, January 2001.
