As enterprises look to ensure more oversight and understanding of their AI deployments, chief AI officers are fast becoming a prominent fixture in the C-suite, charged with decision-making authority over the technology’s sprawling use.
According to IBM research released on Nov. 17, the number of organizations with CAIOs has boomed — rising from 26% in 2025 to 78% as of May this year in the U.K. alone. Yet appointing a CAIO will not, on its own, ensure good governance.
Rhodri Arrowsmith, managing partner at IBM Consulting U.K. & Ireland, said the appointment needs to come with structured governance frameworks and strategy to translate AI’s potential into tangible outcomes.
In this Q&A, Arrowsmith discusses what is driving the rise of the CAIO, the responsibilities that come with the role, and why board-level influence will be essential to making enterprise AI work.
What is propelling the growing adoption of the CAIO role, and what does it bring to an organization that might otherwise be missing?
Arrowsmith: AI is very much at a business-defining moment in terms of providing value, but businesses are struggling to achieve outcomes. Around two-thirds of AI projects are not achieving their intended outcomes, and that’s the kind of situation where a CAIO provides a real opportunity to focus on the architecture, priorities and standards and to decide what needs to happen and how funding is allocated.
Ultimately, it's about creating transparency and a common, metrics-based approach to understanding whether AI investments are delivering value.
What does having a CAIO look like in practice? How can someone in that position gain oversight of AI initiatives across different business functions?
Arrowsmith: Organizations need a standards-led approach to AI implementation, with a governance model created and endorsed at the board level. The CAIO’s role is to act as a sponsor for use cases that will create value and to demonstrate the actual outcomes.
Really, it’s a bit of a flywheel for companies to create much more productivity and better customer experience. You've got a central focal point: someone who can articulate and set the vision for the whole organization and then amplify the stories of what works.
An example I often use is complaints handling. In many organizations, whether that's a bank or a retailer, the process relies on significant amounts of unstructured data. AI can help analyze that information and accelerate the development of recommendations on how a complaint should be handled. That can improve productivity and deliver better customer experience by speeding up responses.
The CAIO's responsibility is to identify those opportunities, sponsor the relevant initiatives and accelerate their development within the context of governance, standards, control, transparency and cost.
How could appointing CAIOs resolve questions of accountability for AI decisions?
Arrowsmith: In terms of decisions around AI, there's quite a lot of variation in where CAIOs sit within organizations, whether they report into data or technology functions, and that can affect responsibility. I think in the long term, the role will move toward the board level, reporting to the CEO or COO.
Ultimately, responsibility for the consequences of AI sits with the CEO, which is why the CAIO needs to operate at that level. Otherwise, the role is constrained in what it can achieve, so business sponsorship is important.
What are the biggest challenges organizations face when introducing a CAIO?
Arrowsmith: A lot of the companies we work with at IBM are large-scale organizations that have grown organically or through acquisitions, creating complicated operating models and proliferation of technologies of different ages.
The challenge for the CAIO is to recognize where value can be achieved within that landscape. The technology and business environment isn't going to change overnight, so the role needs to understand the business as well as the technology, and how value can be delivered within an often complex, dispersed organization.
Embedding all of that understanding within any one individual is, I think, challenging.
How can enterprises address these challenges?
Arrowsmith: That next phase needs to be properly architected, designed and secured. The CAIO can provide an anchor for how that transformation should happen, building on the experimentation and understanding employees have already developed.
Clear boundaries around where AI can act independently and where human involvement is necessary are also crucial. It's important that organizations define the mandate for AI systems, including which actions they are authorized to trigger and where a human must remain in the loop.
That needs to sit within a wider governance framework that establishes transparency and accountability. Organizations need to understand how their AI systems are performing against their objectives and be able to intervene when necessary.
The research highlights why this matters. While 51% of surveyed U.K. CAIOs say AI agents already initiate and execute actions across workflows, 77% say governance, risk or compliance considerations delay AI-driven action. There is clearly a need to reconcile the opportunity to give AI greater autonomy with the controls required to deploy it responsibly.
What would be the first steps for a company looking to integrate this role?
Arrowsmith: The objectives and authority of the role are really important. One thing that struck me in the research was how certain behaviors can improve the success of AI projects. The CAIO needs the ability to pause, adjust and redirect initiatives, decide which workflows AI should be embedded into and have the autonomy to influence different parts of the business with board-level sponsorship. They also need to be able to redirect investment when problems emerge.
Having someone with the experience and understanding to recognize what works, what doesn't and where opportunities lie can make a significant difference.
If I were setting up an organization today, one of the first names on the team sheet would be a CAIO. The next question would be how you give that person the authority and support to deliver value.
Do you expect the CAIO role to become a permanent fixture in the C-suite, and what will distinguish the organizations that get it right?
Arrowsmith: I don't see AI going away or organizations stepping back from it. The question is how it's implemented, which areas are accelerated and how it's controlled.
I expect the CAIO to become as synonymous with the C-suite as established positions such as the chief data officer and CIO. The responsibilities may vary as businesses work out what is most effective for them.
But as more organizations appoint CAIOs, the differentiator won't simply be whether they have someone in that position, but how they've established the role, integrated it into the board and empowered that individual to succeed.
That means putting the right governance framework in place, creating transparency for AI projects, controlling costs and establishing clear performance measures. Organizations also need to react quickly, deciding what to stop, what to accelerate and where to direct additional resources. That's ultimately what will distinguish the businesses that gain a competitive advantage from those that don't.
Editor’s note:This interview was edited for clarity and conciseness.
Scarlett Evans is a freelance writer with a focus on robotics and emerging technologies. Previously, she was assistant editor at IoT World Today, where she specialized in robotics and smart city technologies. She also has a background in the mining and resources sector, with experience at Mine Australia, Mine Technology and Power Technology.