This was one of the most debated questions at our recent executive round tables with more than 20 senior leaders.
Responses ranged widely:
• Some argued ownership sits with technology and data teams.
• Others pointed to business unit leaders and subject matter experts.
• A few said it must sit squarely with the CEO.
What became clear is that in organisations moving quickly along the AI maturity curve, ownership is not a simple “who” – it is a system of aligned responsibilities across the entire organisation.
A few themes emerged:
1. The Board: mandate and guardrails
Boards need more than just awareness of AI. They need:
• A clear productivity and growth agenda that AI is expected to support.
• An appetite to manage AI risk, governance and compliance, not avoid it.
• Visibility of how AI links to strategy, capital allocation and performance.
2. The CEO: aptitude and licence to change
AI is now a leadership capability, not a technical curiosity. Effective CEOs:
• Have enough AI aptitude to ask the right questions and challenge assumptions.
• Hold a clear mandate for change and are prepared to reshape operating models.
• Signal that AI is not an experiment on the side, but central to how the organisation will work.
3. Executives and SMEs: from owners to co-owners
One executive put it well: traditionally a function or project has a single owner with clear accountabilities; AI is a shared responsibility.
That means:
• Technology teams own enablement, architecture, security and integration.
• Business leaders own use cases, value realisation and change in their teams.
• Risk, compliance and legal functions own the frameworks and controls.
• HR and learning own capability-building and workforce impact.
In other words, no single person “owns” AI – but everyone has defined accountabilities within an agreed AI operating model.
The organisations moving fastest are the ones treating AI as a cross-functional capability with shared ownership, rather than a project assigned to one silo.
How is AI ownership structured in your organisation today – and does it match where you want to be on the AI maturity curve?