Why CAIOs need to be a driver—and not coordinator—of change

CAIOs—that’s chief AI officer, for those unfamiliar—are being set up to fail at many organizations, suggests Sebastian Wernicke, partner at finance and economic consultancy Oxera and author of the 2026 book Data Inspired.

Some CAIOs are set up to be drivers of change, while others are stuck as mere coordinators of change, Wernicke said, adding that the coordinators often have limited budgets, small teams, and no reporting line to important business leaders like the CEO.

Yet CAIO coordinators are still expected to be the “center of gravity” for companies’ major AI shifts.

“I think this is where you really see a conflict, because you’re expecting a lot more things out of the role than what it is actually set up to deliver,” Wernicke told IT Brew.

Yet it’s critical for companies to figure out how to optimize the CAIO role. An IBM survey found that more than three in four (76%) global CEOs reported having a chief AI officer in 2026—up from 26% in 2025. With that in mind, Wernicke offers advice for the new class of CAIOs, along with thoughts on how the role might evolve in coming years.

The conversation below has been edited for length and clarity.

What are some “coordination” tasks that you find CAIOs are doing instead of the more strategic work?

The coordination tasks would be if you are, for example, investing a lot of time to educate the entire organization about AI. You’re spending essentially a lot of time building credibility through executive presentations, governance committees, steering boards, and you’re trying to make everybody aware of what AI is. You’re showing the possibilities, which is good…but let’s say you have a production process that you’re going to change with AI. That means it’s not enough to just coordinate. I think you need to go quite deep there and understand: How is the process running today? How could it run in the future? Then, you bring together the different departments; the person responsible for the production itself, and you need to bring in IT; you need to get the budgets right, and so on, and that, from a purely coordinating standpoint, I think is always very hard to do.

What would your advice be to chief AI officers who might feel like they’re being given accountability without authority?

I think it’s to engage with the business, also with the operational parts of the business very early on. There could be a temptation to first bring in that technology, or spend a lot of time building that credibility. Engage with the operational units to understand where their pains in the P&L [profit and loss] actually are. What keeps them up at night? Are they kept up by a need for efficiency? Are they kept up by a need to find the next breakthrough product? The next business model for the business unit? Do they have to de-risk further? Are they kept up by immense complexity? And by understanding these problems, I think that it’s almost inevitable that data and AI will present solutions, but it’s what I like to call a “purpose first, data second,” or “purpose first, AI second” approach, where you really need to understand what the pain points of the business are, and then bring AI to solve those, rather than saying, “I have this technology and I’m now going to bring it to you.”

What’s the solution here? Do you envision a stronger set of requirements for the CAIO?

I think if you establish it, you should be clear from the get-go that this is a transformation role; it may come with a technical label, it may sound like somebody tasked with bringing technology, but it is a business transformation role. You need to have that clarity and also that commitment. One way that I think this commitment is expressed, for example: if the chief AI officer reports directly to the CEO.

If the argument is that chief AI officers are potentially set up to fail, what does failure look like?

With establishing a role like this, of course, companies are looking to create transformative change, and so at some point they will just look at this and say, “We created the role. We made all the investments, do we really see transformative change? Or have we only seen incremental steps?” And when it’s only these incremental steps, that’s, I think, when dissatisfaction starts to happen, and then the role starts being questioned.

We don’t have enough data for the long-term projections, but I think there’s a danger that is somewhat similar to what happened with chief data officers, where among the C-suite they just had the shortest tenures among their colleagues. And that I think also was often attributed to just the value in the end not being clear enough to say, “We’re going to keep this role around.”