Published on 20 July 2026
When the first large language models burst onto the market in 2023, the conversation in boardrooms revolved around efficiency: automating the helpdesk, reducing tickets, speeding up onboarding… Reasonable goals, yes — but ones that barely scratched the surface of what was coming.
Three years on, the reality taking shape is very different. AI has not arrived to make what we already did faster. It has arrived to question why we did it that way in the first place. And that question — uncomfortable, disruptive and necessary — is what is redrawing the IT function from the ground up.
“The mistake is not adopting AI too fast. The mistake is adopting it with the mental maps of traditional IT”
From Reactive IT to Predictive IT
THE OPERATIONAL SHIFT
The classic IT operations model works on a logic of response: something breaks, someone detects it, a ticket is opened, assigned, and resolved. Every step in that chain means lost time and stalled business. AI breaks that cycle at the very first link.
AIOps platforms — Operations Intelligence powered by predictive models — correlate millions of infrastructure signals in real time: logs, metrics, traces, usage patterns. They identify anomalies before they turn into incidents. They automatically resolve part of the operational issues without human intervention.
But the most important leap is not technical; it is cultural. IT teams that have adopted this model no longer talk about “fixing problems”. They talk about preventing states of degradation. It is a paradigm shift comparable to the one manufacturing went through when it moved from corrective maintenance to preventive — and later to predictive.
The New Service Architecture
Where the IT service catalogue used to be a fixed list of standardised offerings, AI turns it into a dynamic, contextual surface. The evolution looks like this:
The IT Technician as a Designer of Intelligent Systems
Perhaps the deepest change is the one affecting people. The dominant narrative speaks of job destruction; experience points in a different direction: a radical redefinition of the value delivered.
The near future IT technician does not configure servers — they design the reasoning flows of the agents that manage them. They do not resolve first line incidents — they define the criteria by which AI decides what to escalate and what not to. They do not document procedures — they train models on the organisation’s tacit knowledge.
“The most strategic role in IT is no longer the systems architect. It is the behaviour architect of intelligent systems”
This demands new skills — understanding language models, prompt engineering applied to operations, algorithmic risk assessment — but also new mindsets: tolerance for ambiguity, the ability to design for emergent behaviours, and the skill to explain AI decisions to non‑technical stakeholders.
The Two Risks No One Is Talking About
WHAT WE SHOULD BE DISCUSSING
1. Governance debt. Many organisations are deploying AI capabilities in IT at a pace that outstrips their risk management maturity. Who decides what an agent can automate autonomously? What audit model applies when an AI makes a production change decision? The lack of clear answers is not harmless: it creates regulatory exposure and erodes internal trust.
2. Operational opacity. Paradoxically, the smarter IT becomes, the harder it is to explain why things happen. Teams that have automated without properly instrumenting their systems discover that, when something fails, they understand less than they did before. AI must not become an operational black box.
What Every CIO Should Be Wondering
Not “How do I incorporate AI into my IT department?” That is the wrong question. The right one is: What kind of internal technology organisation do I want to be in three years’ time — and what role does AI play in that model?
The difference between these questions is the difference between a tactical and a strategic transformation. The first produces pilot projects. The second produces organisations capable of competing in an environment where internal technology is increasingly a real and differentiating competitive advantage.
IT is no longer the department that keeps the infrastructure running so the business can operate. It is the department that defines how artificial intelligence amplifies the capabilities of the entire organisation. That shift in identity is not cosmetic — it is the most important change of the last twenty years.
