Who owns AI is a question every board is now asking, and inside most C-suites the honest answer is that four executives each believe it is them.
In Brief
- AI ownership in the C-suite is not absent — it is contested across four functional claims with no accountability design to resolve them.
- Boards agree the C-suite owns AI (90%), but the C-suite itself splits four ways on who inside it actually does (Pearl Meyer, 2026).
- Each functional executive defines AI ownership from their own remit — technology, data, investment, or operations — which is why the claim fragments.
- Naming the problem as organisational is not the same as designing who owns the fix.
The chief information officer points at the enterprise architecture, the platform and the vendor stack, and reads AI ownership as a technology decision. The chief data officer (CDO) points at the data estate, model provenance and assurance, and reads AI ownership as data governance under a new name. The chief financial officer points at the capital, the productivity case and the assumptions that sit under every business case in the portfolio, and reads AI ownership as an investment discipline. Business unit heads point at the workflows, the customer impact and the P&L that lands on their desk, and read AI ownership as an operations question that only the accountable line executive can answer.
Each of them is correct inside their own remit. That is what makes this difficult to resolve.
AI ownership has been claimed four times, not once
Recent survey work from Pearl Meyer reported by Fortune (Pearl Meyer, 2026) found that 90% of board directors say the C-suite owns AI. Inside the C-suite itself, the answer splits four ways: 32% locate ownership with the C-suite as a group, 22% one level below, 27% with individual business leaders, and 17% with functional heads (Pearl Meyer, 2026). No group commands a majority. Every group commands a plurality of something.
The absence framing does not fit the data. Ownership has already been claimed, and it has been claimed several times, by different executives, each with a defensible reading of their own accountabilities. What the board is looking at is not a vacuum. It is looking at four assertions that cannot all be true at once, and no adjudication mechanism to decide which one is.
The problem is not a governance gap. It is a governance overload. The same decision has been made in four different rooms, and the enterprise has no forum in which those four rooms compare notes.
The four claims persist because each function reads AI through a different accountability
Every function reads AI through the lens it was built to defend. If the CIO’s contract is written against the platform, AI becomes a set of models, integrations and infrastructure decisions, because that is the version of AI the contract covers. The CDO holds the data estate, so the model inputs and the model outputs are both data, and data has always been where the CDO’s accountability sits. Capital allocation is the CFO’s frame, and every AI initiative arrives at the investment committee expressed as spend. The business unit head wears the outcome of the workflow, so AI is either producing a better P&L or it is not.
None of the four is reaching for territory. Each is answering the question in the language of the accountability they already hold. What is missing is the layer above them: the enterprise accountability that decides which slice of the AI question sits inside a given function’s remit, which sits outside it, and where the boundary is drawn. In most organisations, that layer has not been designed. It does not appear by default, and there is no function in the org chart whose job it is to build it.
So the four functional answers behave as though they are competing, when the real design failure is that no one has authorised any of them at the enterprise level. Larridin research reported by BusinessWire in February 2026 found that 58% of executives report no clear ownership of AI inside the enterprise, and 75% still lack a fully implemented AI governance program (Larridin, 2026). Both figures name the same underlying condition: the enterprise-level accountability that would resolve the four functional claims has not been designed. The claims are made; the layer that adjudicates between them is missing.
The board question is about accountability design
When the board asks who owns AI, it is not asking for a name. It is asking whether the executive team has designed an accountability structure it can defend. A CEO who answers with the CIO’s name, or the CDO’s, or a rotating committee’s, is answering a question the board did not ask. The board is asking whether, when the regulatory notice arrives, or the incident briefing lands, or the productivity claim comes due at the investment committee, one decision-maker is accountable and every peer in the C-suite knows the scope of that accountability without having to work it out in the room.
The immediate cost is that the board question cannot be answered coherently. The answer depends on which executive is sitting there when the question is asked. That variation is not evidence of dysfunction in any one function; it is evidence that the design work has not been done at the layer where the four functions meet.
The larger cost is that the design vacuum becomes visible under external pressure. An audit, a regulator, an incident, a public disclosure — one of these will surface the question before the executive team has resolved it internally. By the time it does, the option to design the accountability structure on the executive team’s own terms is gone. What gets designed instead is whatever the external event requires, built by whoever is closest to that event, and the enterprise inherits an accountability model shaped by the crisis. Every quarter the design work stays undone, the gap between what the C-suite intended and what the C-suite ends up with grows harder to close.
The CEO designs the accountability layer
The CEO’s decision here is not to pick a winner among the four claims. Each functional executive is correct about their own remit. Picking one would either overreach that executive’s actual accountability or force the other three to defer on questions they still legitimately own. The decision is to design the enterprise accountability layer that names what sits inside each function’s remit, what sits outside it, and — importantly — what belongs to the layer none of them currently occupies.
That layer is not a committee. Committees exist in most organisations that have a fragmented AI ownership answer, and they are usually the reason the answer stayed fragmented. What is required is a single, accountable owner for the enterprise AI accountability design: the person whose job is to hold the boundaries between the CIO’s remit, the CDO’s remit, the CFO’s remit and the business unit heads’ remits, and to escalate to the CEO only the questions that genuinely sit at the CEO level. That role can be filled by an existing executive whose scope is redefined, by a new role created for the purpose, or by an explicit delegation from the CEO. What it cannot be is a shared responsibility distributed across the four functions that already claim ownership from their own remit, because that is the design that produced the problem.
What this means for senior leaders: The board question about AI ownership is a design question, not a personnel question. The four functional claims will not resolve themselves through goodwill, cross-functional coordination or a steering group with rotating chairs. They resolve when the enterprise accountability layer above them is designed and named. The CEO who does that work before the first external test arrives inherits an accountability structure shaped by strategy. The CEO who waits inherits whichever structure the first regulator, auditor or incident forces on them.
The four answers are all correct at their own level. What remains is the level above them, and it is the only level the CEO can design.
References
- Larridin. (2026, February 3). New study shows C-suite leaders highly confident in AI ROI even as 58% claim there’s no clear ownership of AI and 75% lack AI governance. BusinessWire.
- Pearl Meyer. (2026, April 22). Boards say the C-suite owns the AI strategy. The C-suite doesn’t agree. Fortune.
