The rise and ROI of the chief AI officer

The chief AI officer may be the fastest growing role in business. A new report from IBM’s Institute for Business Value (IBV) that debuted at Think 2026 found that 76% of surveyed organizations said they have a CAIO in 2026, up from just 26% in 2025. And that’s not just at tech companies like Meta and Salesforce, but also at enterprises like Heineken, WPP, Nike and CVS Health. That kind of growth can look like froth. But the IBV found that companies with a chief AI officer had a 5% higher return on their AI investments. How did that happen?

Just like the technology that drove it, the CAIO role has matured over the past few years. “It used to be that chief AI officers were more figureheads—AI evangelists promoting AI,” said Jacob Dencik, Research Director at IBV in an interview with IBM Think. “But now they’re actually driving real transformation with AI and helping enterprises move from pilots to wide-scale implementation.”

Schneider Electric is a clear example of the focus on implementation. The global energy technology company created its chief AI officer role in 2021, well before generative AI forced the issue for much of the corporate world. From the outset, Schneider’s mandate for AI leadership was less about experimentation for its own sake and more about operational impact. As Philippe Rambach, the company’s Chief AI Officer, emphasized in his conversation with IBM Think, AI at Schneider “always starts with a business need, not the technology.”

Still, not everyone is convinced companies need the role. Tim Crawford, Founder and CIO Strategic Advisor at research firm AVOA, sees the CAIO moment as familiar. He compares it to the rise of the chief digital officer a decade ago—a role that often emerged quickly at companies, with mixed results, especially when organizations rushed to bring in external CDOs, he noted. These leaders sometimes struggled because they “weren’t as in tune with the business”; meanwhile, the core responsibility of digital transformation often fell back on the CIO.

On paper, the rise of the chief AI officer can seem redundant. Many companies have added AI to the remit of CIOs, CTOs, chief data officers and chief digital officers. But as organizations have discovered over the past two years, AI doesn’t behave like earlier technologies. Its reach is broader, its pace faster and the expectations around it far higher.

“AI is crossing the entire enterprise in a way that most other technologies haven’t,” said Dencik. Unlike cloud computing or enterprise software, which largely lived inside IT, AI is something “every C-suite member and every employee potentially has an expectation about,” he said, including “what it should be doing and how fast it should be delivering value.”

At Schneider Electric, that breadth helped shape the role early on. Rather than positioning AI as a standalone technical function, the company organized its efforts around a hub-and-spoke model: a central AI team responsible for strategy, standards and tooling, paired with execution embedded in business units. The approach, Rambach said, helps keep AI close to real operational problems while avoiding fragmentation and duplicated effort.

Crawford agreed that AI fragmentation is a top challenge for leaders, but he cautioned against assuming every organization needs a standalone CAIO to address this challenge. In some cases, he said, the responsibility for AI can sit with the CIO or even the CEO, provided there is clear accountability and strong cross-functional coordination. Other companies, from SAP to Nike, have combined the CAIO role with another C-suite title. Philip Herzig at SAP, for example, is both CAIO and CTO, and Alan John at Nike is Global Head of Data and AI. The larger issue, said Crawford, is not which executive owns AI, but whether an organization has the leadership structures needed to guide it responsibly as it spreads across the business.

Furthermore, Dencik pointed to a growing gap between ambition and reality: companies are investing heavily in AI, yet sometimes struggling to move beyond pilots or measure impact. The result, he said, is pressure to create “a dedicated capability in the organization that can actually convert AI into business value,” rather than letting initiatives sprawl across departments without coordination. Hence, they hire a CAIO.

That need for central authority helps explain why many chief AI officers don’t necessarily report to technology leadership. According to IBM’s research, a significant number report directly to the CEO or even the board, underscoring that AI is increasingly treated as a strategic business issue rather than a back-office concern.

“It’s about bridging the business side of the house with the technology side,” Dencik said. “AI is no longer just a tech discussion.”

Schneider Electric took a similar view when it established the role, deliberately anchoring AI leadership in business priorities such as energy efficiency, supply chain optimization and sustainability outcomes. The chief AI officer’s remit, Rambach said, is not to choose models or tools first, but to ensure AI investments are tightly aligned with where the business needs measurable improvement.

Even in companies with dedicated CAIOs, those leaders do not always function as peers to other C-suite roles. In practice, Crawford said, many AI leaders operate more effectively as SVPs or heads of horizontal functions—roles empowered to coordinate and govern, rather than compete with CIOs, CFOs or COOs. What matters most, he said, is not the title, but the mandate: the ability to bring the right people together to set priorities, run experiments and put guardrails in place.

That model is increasingly common, often formalized as an AI council. In Crawford’s view, such councils should be cross-functional by design, focused on outcomes rather than optics. “Companies shouldn’t play the CAIO as a marketing ploy,” he said. “Customers don’t really care whether you’re using AI. They care about the result.”

Despite the explosion in titles, the chief AI officer remains widely misunderstood. It is not, as some skeptics joke, a “Chief ChatGPT Officer,” responsible for sprinkling generative AI across slide decks. Nor is it meant to be a compliance-only role, fixated on regulation while the rest of the business charges ahead.

In practice, the job sits somewhere in between. As companies race to adopt AI, many have discovered that hype can damage your credibility as much as inaction. Daniel Hulme, Chief AI Officer at WPP, has been blunt about that risk. “Every time there’s a new technology that comes along, people get very excited,” he said. “And then they apply those technologies to solving the wrong problems.”

The importance of staying grounded is something Schneider’s AI leadership has reinforced repeatedly. Starting with business needs, the company has argued, helps cut through the noise—forcing teams to justify AI investments by impact rather than novelty.

Crawford echoed that concern, arguing that much of today’s competitive posturing amounts to “AI washing.” From a customer view, Crawford said, value creation matters—but so does trust. Customers may welcome smarter products and services, but misuse of data can quickly erodes any goodwill. The lesson for AI leaders, he said, is that success is measured less by technical sophistication than by whether AI meaningfully improves how a business serves its customers.

As companies seek candidates with a mix of business, technology and transformation experience, leaders disagree on whether an AI lead should come from within or outside the company. One camp argues that the most effective chief AI officers are promoted from within, since AI transformation depends less on technical novelty than on deep institutional knowledge.

At Schneider, which established its CAIO role early, elevating an internal leader helped consolidate the company’s already-active AI efforts across energy management, industrial automation and sustainability. That insider perspective made it easier to align AI initiatives with real operational priorities and introduce governance without slowing momentum. For supporters of this approach, the CAIO’s primary task is integrating AI into how the business already works, so the best candidate is an experienced insider.

Others see clear advantages in hiring externally, particularly when a company wants to accelerate change or rethink longstanding assumptions. Heineken’s decision to bring in Surajeet Ghosh from outside, for example, reflected a desire to build AI capabilities almost from scratch. With experience applying AI across industries, Ghosh brought in a data-led approach to make key decisions across the brewer’s global value chain, a topic he discussed in depth in a recent episode of the Smart Talks podcast with Malcolm Gladwell. Advocates of external hiring argue that outsiders can help organizations avoid incrementalism, impose clearer ROI discipline and drive faster movement from pilots to production—even if they face a steeper learning curve around culture.

In the Smart Talks conversation with Gladwell, Ghosh shared a telling example: soon after starting as Heineken’s Chief AI Officer, he helped analyze the ROI on Instagram advertising for Heineken versus Dos Equis. After applying AI to this challenge, they were able to improve sales by 30%.

When it comes to sourcing a CAIO, a third view is increasingly taking hold, with experts arguing that the internal-versus‑external debate matters less than how the role is defined. Across companies making real progress with AI, structure beats labels. If organizations with a formal CAIO role report better outcomes, it’s not because of the title; the differentiator is how AI is organized and governed.

From researchers to CxOs, there is a consensus that no chief AI officer succeeds without strong experience in technology, business strategy and change management. These capabilities can be developed internally or hired from outside, but they must be paired with organizational influence and narrative clarity.

“No one person should own AI—it has to be shepherded,” said Lula Mohanty, a Managing Partner for IBM Consulting in the Middle East and one of the authors of IBM’s 2025 Chief AI Officer report, in an interview with IBM Think. The CAIO, said Mohanty, is ultimately an orchestrator of change, focused on adoption, outcomes and embedding AI into the organization.

IBM has no chief AI officer. But several IBMers cited Joanne Wright, the company’s SVP of Transformation and Operations, as the equivalent. Wright sits at the intersection of every operational domain of the company, giving her a unique vantage point to see exactly where AI should be deployed, and when. When asked whether she’s a de facto CAIO, Wright told IBM Think, “Yes and no.”

She elaborated: “I am accountable for how we use AI to fundamentally change how the company works. This means how we serve clients, work with partners, manage our workforce and automate processes that used to require enormous amounts of human effort. But I don’t ‘own’ AI.” Instead, she said, each leader at IBM is accountable for AI adoption on their own teams, and her role is to remove any friction. “What I am proud of is that we have built an operating model where AI strategy is so deeply embedded into the business strategy,” said Wright. “Without a central orchestrator, your business ends up with solutions that don’t talk to each other, are fragmented and can’t effectively scale.”

At Schneider Electric, where the hub-and-spoke operating model has been central to the organization’s AI efforts, AI teams have moved quickly while maintaining common standards, governance and guardrails. Five years into the role, the company’s CAIO brings something many newer CAIOs lack: evidence of what it takes to sustain AI at scale over time.

WPP, by contrast, approaches AI through the lens of its marketing supply chain, layering generative models onto an already well-defined operational framework. “Those other things didn’t change,” Hulme said. “It just heightened and enhanced what we were already doing.”

That emphasis on operating models aligns closely with Crawford’s view. The real work of AI leadership, he said, is ensuring the organization is focused on the right priorities, experimenting quickly and scaling what works—without “running with scissors” when it comes to governance and risk.

As companies move to scale agents, governance is becoming increasingly critical. As important as having someone function as AI orchestrator is pairing that role with orchestration-led governance. This approach changes where control lives. Instead of writing policies and hoping teams interpret them correctly, organizations are embedding guardrails directly into the systems running AI, spanning models, agents and increasingly complex agentic ecosystems.

The payoff is already tangible. Consider that nearly seven in 10 executives now admit they lack full visibility into the AI their teams are using—or even where those systems operate. And yet, companies that adopt orchestration-led governance tend to close that gap quickly: they are more than twice as likely to have full visibility into AI assets, 169% more likely to maintain transparent documentation and 132% more likely to protect data through anonymization, impact assessments and strict access controls.

Crucially, this added control does not come at the expense of performance. These 2,000 organizations surveyed report 29% lower losses from AI irregularities and achieve 20% higher ROI, alongside stronger gains in productivity and revenue.

Together, those outcomes underscore why the chief AI officer role may continue to evolve away from ownership and toward coordination, said IBM’s Dencik. The CAIO’s effectiveness increasingly depends not on building models or drafting policy, but on orchestrating systems where innovation and governance advance together by design. In an era where AI expands faster than any single leader can oversee manually, orchestration-led governance is what makes that leadership durable—and what ultimately separates scalable AI from unsustainable ambition.