CAIOs drive better AI results, but companies still lag on AI leadership

Scattershot approaches to AI adoption aren’t providing meaningful returns, but for many CIOs, that message still isn’t landing.

Over the past two and a half years, many enterprises have launched dozens of AI projects without creating long-term strategies, resulting in huge numbers of failed proof-of-concept experiments. Meanwhile, full-fledged AI strategies, along with dedicated AI leadership, not only lead to immediate impacts but also drive success over the long term, two new studies suggest. Still, most organizations aren’t taking the necessary steps, such as designating an AI point person, some experts say.

According to a recent Gartner survey, 45% of leaders at organizations with high AI maturity say their AI initiatives remain in production for three or more years, ensuring long-term impact and value. That’s compared to only 20% of leaders at organizations with low AI maturity saying the same about their AI project survival rates.

Gartner defines AI maturity using a number of factors. High-maturity organizations regularly measure the benefits of their AI initiatives using multiple metrics, the analyst firm says, and most have centralized their AI strategy, governance, data, and infrastructure capabilities. Most AI mature organizations also run financial analysis on risk factors, conduct ROI analysis, and measure customer impact.

AI leaders needed

Beyond components like strategy and governance, leadership matters. More than nine in 10 organizations that meet Gartner’s high AI maturity definition have appointed a dedicated AI leader, whether that be a chief AI officer or a CIO, CTO, or other leader with AI responsibilities.

In a related study, the IBM Institute for Business Value finds that organizations with CAIOs achieve 10% greater ROI on AI spending and are 24% more likely to say they outperform their peers on innovation.

IBM reached out to about 2,300 organizations, 26% of which had a CAIO, up from 11% in 2023. More than half of CAIOs come from the organizations’ internal talent pools, and two-thirds of CAIOs surveyed expect that most organizations will have the position within the next two years.

There seems to be a disconnect between the need for AI leadership and actual appointments of CAIOs, or at least giving the CIO more AI authority, says Matthew Candy, global managing partner for generative AI at IBM Consulting. More than half of organizations appear to be still in the AI pilot phase, he says.

“You’ve got organizations still stuck in this kind of pilot purgatory,” he says. “They’re still in this test/learn/pilot activity, and this role is key to getting them out of that state of piloting into execution at scale.”

The AI leader within an enterprise needs budget and decision-making authority to move pilots into production, Candy adds. However, the role shouldn’t be all about the technology; one of the important goals for a CAIO or a CIO with AI responsibilities is change management and employee buy-in.

“It’s far broader than just technology adoption and deployment,” he says. “This is about how you orchestrate the company’s business transformation, and how you embed AI into operations, how you establish governance, how you upskill people and teams, drive a cultural shift, and take the whole organization on that journey to get to a place where you can measure the impact at scale.”

Building maturity

With all these factors to consider, AI maturity doesn’t happen overnight, says Birgi Tamersoy, a senior director analyst at Gartner. But enterprises can take small steps toward better AI outcomes, he adds, by creating mature AI strategies through trial and error, he adds.

“In the beginning of your AI journey, you lack a formal strategy,” he says. “As you go along, you create your AI strategy, then you deeply integrate it into your business — then your AI strategy starts potentially shaping your business strategy as you get more mature.”

Maturity goes beyond strategy, however, and is multidimensional, he adds. “You have to know what you are doing in the engineering side of things, and you have to have the governance processes in place to increase the trust, security, and safety of your solution,” he says. “You start with experimenting a little bit. And as you gain more experience, and as you gain more knowledge, then you put more mature processes in place.”

Starting small can be a good approach, and organizations should focus on moving from being operational-level AI users to transformational-level, adds David Gucker, COO at cloud infrastructure provider Vultr. This change requires a shift away from experimentation to full-scale integration of AI.

“Most organizations start with isolated use cases or POCs,” he adds. “But as these initiatives demonstrate value, the next step is to scale AI across multiple departments and functions.”

To reach maturity, enterprises should ensure their leadership teams are in synch on a shared vision for how AI will create business transformation and measurable value, Gucker adds. Dedicated centers of excellence can drive AI strategies forward.

“If this is too overwhelming, setting up an internal committee dedicated to AI is also an option,” he adds. “This cross-functional team should include both executives and technical leads who can act as champions of AI adoption while ensuring governance, security, and compliance. In either case, a named individual as the owner of the AI domain is necessary to ensure the project maintains momentum.”