Enterprise AI is at a pivotal juncture where CIOs must gauge the automation and productivity benefits AI can bring while demonstrating how AI delivers ROI and business value to their organizations. A fundamental requirement of the CIO job in 2026 will be to determine when launching new AI technology stops being additive to an organization’s goal of establishing efficient workflows and reducing costs, explained Jay Upchurch, CIO for BI company SAS .
This is the year in which CIOs will transition into the role of “Chief Integration Officers,” identifying how to integrate existing AI technologies within their organizations instead of continually launching new AI systems, predicts Upchurch.
“The CIO role, especially in ’26, is going to have to be much more rooted in integration — integration between [AI] agents and systems … and then making sure that we’re doing that in a secure and governed way,” Upchurch said.
(Source: Shane Snider/Data Center Knowledge) SAS CIO Jay Upchurch in a conversation with InformationWeek’s Kelsey Ziser.
Over the past two years, CIOs have been pressured to implement Gen AI — and later agentic AI — as fast as possible, said Upchurch. However, demonstrating ROI from launching these AI technologies hasn’t been easy, creating a “cost-efficiency dilemma for all CIOs,” he said.
He likens the difficulty in translating AI deployments into ROI to the early days of cloud computing and virtualization. During the cloud transition, CIOs “had to refactor all our workloads and reimagine how we ran with that,” he said. “It’s the same way with AI — AI for automation’s sake is not enough.”
The focus now should be on how to integrate existing AI investments.
Managing the AI integration process
Foundation piece. The process for integrating AI technologies begins with identifying the “agentic foundation” that an organization is going to use, according to Upchurch. For example, SAS is a “Microsoft shop,” he said, so it’s using Microsoft Copilot as the foundational piece to ensure compatibility across existing productivity tools. “We need that digital assistant that is a natural to our flow of work.”
Regardless of which vendor a CIO selects for the organization’s main AI platform, it’s important to have a foundation that works across multiple IT systems, he underscored.
Integration decisions: Build vs. buy. Once a CIO has selected the foundational AI system, they can get to work on integrating that foundational AI with other AI technologies that the organization has deployed. But CIOs should keep in mind that this process will take time because AI technologies are still evolving, Upchurch said.
“I may have an enterprise system that doesn’t have an agent yet that I can communicate with,” he said. “That doesn’t mean I should rush and build one. I should be patient enough that I pay money to that enterprise software vendor to let them show up with the agent that I can then integrate in with. I don’t have to build it all myself.”
Security and governance. The third component that CIOs should focus on during the AI integration process is security, data and access, Upchurch said. CIOs will need to ensure that employees have access to only data that’s relevant to their role.
“So when that agentic integration is happening, it’s knowing what roles are, what [employees] are entitled to know from a data standpoint — making sure that label on all data and the other systems is consistent so you’re not seeing things you shouldn’t,” he said.
