There is a popular narrative doing the rounds right now that says the world is splitting into two groups: people who adopt AI tools and people who don't. Those who adopt will thrive. Everyone else gets left behind. Permanently.
Parts of it are true. But it misunderstands where the real divide is forming.
The gap that matters isn't between individuals who use AI and individuals who don't. Within three years, everyone will use AI tools whether they realise it or not. They will be embedded in every piece of software we touch. That part resolves itself.
The gap that matters is at the business level. And it is forming right now.
The Curve Is Not a Straight Line
AI capability from 2022 to 2035 follows an S-curve. Understanding where we sit on that curve changes how you think about everything.
From mid-2022 to early 2025, AI went from a novelty that couldn't do basic arithmetic to systems that pass professional exams, write production code and reason through complex multi-step problems. That was the steep ramp. Exponential growth from a low base. Impressive, but the starting point was so low that it was easy for senior leaders to file it under "interesting but not yet relevant to my business."
That was a reasonable position in 2024. It is not a reasonable position now.
From 2026 to 2029, we enter what I think of as the capability explosion. Raw benchmark performance begins to plateau but applied capability, what these systems can actually do in the real world, accelerates sharply. Autonomous agents operating software without human intervention. Multi-agent systems coordinating complex workflows. AI that doesn't just answer questions but executes entire workstreams.
This month alone, Codex, Opus 4.6 and OpenClaw landed within weeks of each other. The people building these systems are openly saying they are no longer needed for the core technical work of their own jobs.
We are at the inflection point. The steepest part of the curve is happening now and over the next two to three years.
Three Categories of Organisation Are Forming
This acceleration is creating three distinct categories of business. The competitive dynamics between them are already visible and they will intensify sharply.
AI-native entrants
We are only just starting to see these appear, but the wave is coming. These are businesses built on AI from day one. Their first hires are agents, not people. Humans provide judgement, direction and relationship management. Everything else is automated. They carry no legacy systems, no legacy processes, no legacy cost structures.
A five-person AI-native firm can compete with a fifty-person traditional operation in many knowledge-work domains. Not in every domain and not yet across the board, but in enough areas to be profoundly disruptive.
AI-augmented incumbents
Existing organisations actively transforming how they work. Redesigning workflows, retraining teams, deploying AI across operations, rethinking their value chains. These businesses can absolutely compete. But the transformation requires exceptional leadership capability.
You need people who can drive change at pace while managing the human complexity of workforce restructuring. People who can evaluate where AI adds genuine value versus where it introduces risk. People who can maintain organisational cohesion while fundamentally changing how work gets done. That combination is rare and most organisations have not invested in building it.
AI-resistant organisations
Not because they consciously rejected AI. Because they moved too slowly, underinvested, treated it as an IT project or could not overcome internal resistance. They ran a few pilots, delegated it to a working group and called it progress.
These businesses will not fail overnight. But they will watch their margins compress quarter by quarter as competitors who moved faster pull further ahead.
Why the Gap Becomes Permanent
This is where most analysis stops. But the critical insight is not that these three tiers exist. It is that the gap between them compounds.
The AI-native competitor generates higher margins from a lower cost base. That surplus gets reinvested into better AI capability, not legacy overhead. Better capability attracts stronger talent because the best people want to work with cutting-edge tools, not fight legacy systems. Stronger talent produces faster iteration. Products and services improve more quickly. The lead grows.
Meanwhile the traditional competitor is trying to fund a transformation from shrinking margins while simultaneously trying to retain people who can see the writing on the wall.
There comes a point where the cost of catching up exceeds what the business can afford. The window closes.
We saw exactly this dynamic play out with digital disruption in the 2010s. The difference is the timeline. Digital disruption played out over a decade. AI disruption is playing out over two to three years.
Think about what it means when someone can stand up a business with AI agents handling customer service, content creation, data analysis, financial modelling, market research and routine operations from day one. Their fixed cost base is a fraction of an incumbent's. Their speed to market is dramatically faster. Their ability to iterate and pivot is almost frictionless because they are not managing organisational change, they are reconfiguring systems.
The Questions That Matter Now
If you lead a team, a function or a business, there are four questions worth sitting with.
What does your business look like if an AI-native competitor enters your market tomorrow with 10% of your headcount and none of your overhead?
Which parts of your value chain are genuinely defensible and which are vulnerable to automation?
What is your organisation's competitive moat when execution capability becomes commoditised?
How fast can you realistically transform your operating model before margin compression makes transformation unaffordable?
These are judgement questions, not tool questions. They require a combination of AI literacy, strategic thinking, comfort with ambiguity and the willingness to make uncomfortable decisions about your own organisation. No amount of prompt engineering courses will help you answer them.
Where This Lands
The premium in the years ahead will not be on leaders who can use AI. That will be table stakes.
It will be on leaders who can think clearly about what AI changes for their industry, their workforce, their competitive landscape and their decision-making process, and who can then execute on that thinking with the people around them.
Judgement under pressure. Influence without authority. Sense-making in ambiguity. The nerve to lead through disruption. Execution discipline.
These are not soft skills anymore. They are the hard skills of the AI era.
And the window to build them is narrow and closing.
If you want a clear picture of where you stand across judgement, execution, influence and future readiness, start with our free capability assessment.
Build the capability before the window narrows.