Enterprise leaders are no longer asking whether AI matters. That conversation is over. Spanning marketing to supply chain, clinical diagnostics to call centres, the impact of AI is now material, measurable and multiplying.
At a recent cross-sector roundtable with senior leaders, the mood had shifted. Instead of AIs brilliance, the quiet, systemic challenges of making it work was at the forefront. It was clear: AI is not the bottleneck. Organisations are.
This shouldn’t surprise us. Every major wave of transformation, cloud, mobile, DevOps, agile, has eventually hit the same wall: not capability, but integration. Not feasibility, but ownership. And now, with AI, we’re feeling it again, only faster, and with more at stake.
The myth of readiness
Almost every enterprise today has “AI Readiness” Matrices. However, true readiness is rarely as linear as governance, infrastructure and use cases. Beneath the surface lies decentralised ownership, legacy architecture, shadow IT, overlapping committees and misaligned incentives. Data is everywhere and nowhere, Procurement isn’t built to buy things that learn, and Legal is often seen as a blocker.
There’s no shortage of AI ambition, but few are willing to own the integration debt that real implementation creates.
Small wins, big friction
AI’s most visible progress is happening at the micro level: accelerating internal workflows, summarising meetings, classifying documents, managing overflow in customer contact centres. Whilst welcome, on account of their value and demonstration of real-world AI behaviour, these modest wins often come with disproportionate effort due to organisational friction.
AI reveals your real structure
One of the most striking insights was that AI behaves like an organisational diagnostic. At scale, it reveals the true structure beyond the PowerPoint deck; the one people experience daily – who can make decisions without escalation? Who can access data? How fast can insights flow between teams? Where does trust live, and how is it earned?
AI rewards flow, not control, and many enterprises are not built for flow.
This is where many AI initiatives quietly stall; not because they lack value, but because they demand a level of coherence the organisation doesn’t yet possess.
Strategy is being rewritten in real time
One of the quieter tensions is strategic. CIOs are caught between the board's demand for a coherent AI strategy and the experimental, fragmented reality.
The emerging strategy is adaptive and shaped by real use-case feedback rather than long-term roadmaps. This isn’t weakness, it’s wisdom. But it also creates discomfort. It runs against how many organisations are wired to operate: with top-down certainty, fixed KPIs, and quarterly targets.
The next frontier isn’t capability: it’s coherence
As technology gets better, organisational coherence must be deliberately built, cross-functionally, and often in tension with current structures.
Coherence doesn’t mean uniformity. It means knowing where you’re headed, aligning on principles, and building just enough structure to move at pace. It means ensuring that experiments in one part of the organisation aren’t quietly erased by inertia in another.
That’s not a technology play: it’s a leadership one.
The CIO’s challenge is beyond AI deployment, it’s to create conditions where AI can thrive, experiments can scale, learning can spread, and human trust can grow alongside machine output.
Final thoughts
Whilst our CIO audience didn’t end in consensus, they did end in clarity about the present state of enterprise readiness. There was a shared truth; AI introduces a mirror that reflects organisational design, culture, habits and biases – in real time and at scale.
If we address the fundamentals, the path forward becomes clearer. Otherwise, we’ll keep building systems that are intelligent in theory but incoherent in practice.
And no model – no matter how advanced – can solve that for us.
