Gartner: CIOs must balance AI value with risk — and fast
A useful leadership test is to separate autonomy from automation. Not every workflow needs an agentic model, and overusing AI can create cost, complexity, and weak benefits realization. CIOs should require each initiative to show a measurable business outcome, a bounded risk profile, and a clear owner before it moves beyond pilot. That shifts the conversation from enthusiasm and tooling to portfolio discipline.
The workforce implication is equally material. AI value is more durable when leaders redesign work, not just remove tasks. Reskilling efforts can turn efficiency gains into revenue, service quality, or capacity gains, but only if HR, business leaders, and IT jointly plan role transitions and adoption support. Otherwise, shadow AI and change resistance will grow in parallel with spending.
Practical next questions for leadership teams include:
- Which AI use cases truly require autonomy, and which should remain conventional automation?
- Do we have a single system of record for AI approvals, controls, incidents, and model accountability?
- What value metrics will justify scaling, pausing, or stopping each initiative?
- What reskilling commitments are needed so productivity gains translate into business outcomes?
ORLANDO, FLA. — More than four years into the generative AI boom, spend is growing at an unprecedented rate, while proven successes continue to be a challenge.
Gartner reported that spending on AI has increased 18-fold over the past two years, while only 13% of CIOs have realized significant or transformational value from AI, according to Gartner’s annual CIO survey released Monday, based on responses from more than 2,000 IT executives.
But the spend-success rubric is more multifaceted than project investment. CIOs also have to determine the enterprise thresholds for risk and safety.
Indeed, only 21% of CIOs have agentic AI governance today, but 71% will deploy agentic AI within 12 months, said Kristin Moyer, a distinguished vice president analyst at Gartner.
“That is out of alignment,” she added. “It’s not possible to create value faster than AI can create risk in that situation.”
Given that the two are intricately connected, Gartner advised CIOs to utilize a value-risk flywheel. Daryl Plummer, vice president, distinguished analyst and chief of research at the consultancy, likened the flywheel to a racecar that has to balance speed with drag, or governance. Finding and maintaining balance can create steady momentum that propels rapid adoption while managing risks.
Moyer and Plummer explained how the flywheel can be applied to facets of AI management, including reimaging work, reskilling employees, reducing AI waste, managing shadow AI and establishing proper governance.
Plummer also pointed to Ikea as an example of a company that has managed to balance the risk and value of AI by reskilling 8,500 call center workers to sell interior design while chatbots took over routine calls. This effort brought in more than $200 million in additional revenue in fiscal 2025.
Achieving alignment between the risk and reward of AI means identifying the business outcomes of AI deployments and taking a proactive approach to secure the use of AI within organizations, the analysts said.
CIOs need to take a proactive approach to limiting “careless consumption” of AI and monitoring AI misuse to combat shadow AI. Establishing an internal AI control team will also be key to monitoring AI, according to Gartner. AI insurance is valuable to have when those efforts fail, but it can’t be the only solution. Indeed, only 1% of CIOs have purchased AI insurance but 34% are thinking about it, according to Gartner.
Since AI governance is often fragmented, Gartner recommends establishing an AI governance platform that becomes a system of record for safety and governance. CIOs should work closely with their organizations’ CFO and legal teams to manage that platform.
Moving forward, before increasing investment in any AI initiatives, CIOs need to demonstrate measurable business value, Moyer advised.
“The biggest drag on AI momentum is using it when you don’t need to,” Moyer said. “Use AI agents when you need autonomy, not just automation.”
How is your organization balancing the pursuit of AI value with managing AI risk? Let us know at [email protected].
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.