Blog Details

Ask most companies who owns AI and they point at the IT department. That single reflex is why so much AI spend produces so little. It quietly reframes a business decision as a technical chore, and technical chores do not change how a company makes money.

AI strategy should sit with the CEO and the senior leadership team, not the IT department. AI is a decision about where the business competes, what it automates, and what it stops doing, and those are ownership questions, not tooling questions. IT is essential to delivery, but when it owns the strategy, ambition shrinks to what fits the current systems.

Why the default answer is IT, and why that is a trap

The instinct is understandable. AI feels technical. It involves models, data, and infrastructure, all of which live in IT. So leaders hand it over, relieved to have found an owner, and move on.

The trouble is what gets lost in that handoff. IT is measured on delivery, stability, and cost control. Given AI, a good IT team will do exactly what it is rewarded for: find a safe, contained, technically sound use case and ship it reliably. That is competent. It is also small.

The questions that make AI valuable are not on IT's scorecard. Which parts of our business model are exposed to AI-enabled competitors? Which processes should we stop doing entirely rather than automate? What are we willing to change about how we serve customers? Those are chief-executive questions, and no department can answer them upward.

Strategy is about trade-offs only the top can make

A real AI strategy is a series of trade-offs. It means choosing to invest here and not there. It means accepting short-term disruption to a workflow for a long-term gain. Sometimes it means cannibalising a comfortable revenue line before a competitor does it for you.

Those decisions cross every function. They pit the interests of one department against another. They require someone with authority over the whole business to say this matters more than that. An IT leader, however capable, does not have that mandate, and should not be asked to fake it.

When AI strategy sits below the level that can make company-wide trade-offs, it stops being strategy and becomes a shopping list.

What happens when IT owns the strategy

The failure mode is predictable and polite. Nothing blows up. The AI programme simply underdelivers in a way that is hard to pin down.

• Use cases are chosen for technical convenience, not business impact.
• Projects that would require other departments to change get quietly avoided.
• Ambition is capped at what the current data and systems already support.
• The programme optimises for not breaking anything, which rules out the bold moves that create value.
• When results are modest, leadership concludes AI is overhyped, when in fact it was under-sponsored.

This is how a company can spend real money on AI, tick the box, and still be outmanoeuvred by a competitor who treated it as a leadership priority.

So what is IT's actual job here?

This is not an argument for sidelining IT. The opposite. IT is indispensable, but to delivery, not direction.

Once leadership sets the strategy, deciding which problems matter and what winning looks like, IT owns the how. Architecture, security, data pipelines, integration, reliability, and the unglamorous work of keeping systems running are exactly where IT should lead, and where the strategy lives or dies in practice. A brilliant strategy with weak delivery is just an expensive opinion.

The healthy split is simple. Leadership owns what and why. IT owns how and how well. Blur that line by pushing strategy down to IT, and you get small results. Blur it by having executives dictate architecture, and you get fragile systems. Both roles matter, in their own lane.

The CEO does not need to be technical

A common objection is that the CEO does not understand the technology well enough to own the strategy. This misunderstands the job. The CEO does not need to know how a model is trained any more than they need to know how the accounting software calculates tax.

What the CEO must own is the business logic: which problems are worth solving, what the company is willing to change, and what a win is worth. The technical depth comes from the team and from advisers. The judgement about where the business should place its bets cannot be delegated, because no one else can see the whole board.

This is often where an external partner helps most, translating between the executive view and the technical one so that AI strategy and advisory connects boardroom ambition to what is genuinely buildable. Across more than 670 projects, the engagements that compound are the ones where a senior sponsor owned the outcome, not the ones handed to a technical team to figure out alone.

How to fix ownership without a reorganisation

You do not need a new department. You need three things. First, a named executive sponsor who owns AI outcomes and reports at board level. Second, a short, ranked list of business problems that sponsor cares about, set before any tool is chosen. Third, a clear delivery mandate for IT to build against that list to a high standard. That is the whole fix, and it changes results more than any single tool ever will.

The tell-tale signs your ownership is wrong

You can usually diagnose misplaced ownership without an audit. A few signs give it away.

• When you ask who owns AI, the answer is a department, not a person.
• The AI roadmap is a list of tools and integrations, not a list of business outcomes.
• Every proposed use case can be delivered without any other department having to change.
• Progress is reported in features shipped rather than numbers moved.
• The board hears about AI as a technology update, not as a competitive question.

None of these mean IT is doing a bad job. They mean IT was handed a job that was never theirs to own. The fix is not to criticise the technical team. It is to put the strategy back where the trade-offs can actually be made.

Why this gets harder, not easier, as AI improves

It would be comforting to think this is a transitional problem that fades as the technology matures. The opposite is true. As AI gets more capable, the decisions it forces get bigger, not smaller. Whether to automate a function, restructure a team, or change a service model are not questions that shrink with better tooling. They grow.

That means the cost of parking AI strategy in IT rises over time. Early on, when AI handled only narrow tasks, treating it as a technical project did limited damage. As it starts to touch pricing, staffing, and the shape of the business itself, leaving those calls to a delivery function becomes a serious strategic risk. The companies that get ownership right now are buying themselves an advantage that compounds precisely because their rivals keep filing AI under information technology.

The bottom line

Treating AI as an IT project is the most common way to cap its value. IT should absolutely own delivery, security, and the technical craft that makes AI reliable. But the strategy, the choice of where to compete and what to change, belongs with the CEO and senior leadership, because only they can make the trade-offs that turn AI from a tidy experiment into a business advantage.

Frequently Asked Questions

Who should own AI strategy in a company?

The CEO and senior leadership team. AI strategy is a set of business trade-offs about where to compete and what to change, which only executives with a company-wide mandate can make. IT owns delivery, but strategy sits at the top.

Does the CEO need to be technical to own AI strategy?

No. The CEO owns the business logic: which problems matter, what the company will change, and what a win is worth. Technical depth comes from the team and advisers. The judgement about where to place bets cannot be delegated.

What is IT's role if it does not own AI strategy?

Delivery and craft. Once leadership sets the direction, IT owns architecture, security, data pipelines, integration, and reliability. That is where strategy succeeds or fails in practice, and where IT should clearly lead.

How do I move AI ownership up without a reorganisation?

You do not need a new department. Name one executive sponsor who owns AI outcomes at board level, give them a short ranked list of business problems to solve before any tool is chosen, and hand IT a clear mandate to build against that list well. Those three moves change results more than any single tool.

If AI in your company has quietly become an IT project and the results feel capped, an outside view can help you reset ownership and rank the problems worth solving. Book a free AI opportunity assessment through our contact page for an honest, no-obligation read on where AI should sit and what to prioritise.