Today’s higher ed IT leaders must strike a delicate balance between enabling AI innovation and maintaining a reliable, secure digital backbone.
Balancing AI innovation with Core Systems
“This is an absolutely game-changing time in higher ed IT,” says Ed Hudson, CIO at the University of Kansas. “There is no part of our university environment that isn’t being touched by AI in some way.”
In the past, Hudson says, faculty and researchers would look to their institutions’ IT teams for tech solutions. Now, the academics are “light-years ahead of me” in their grasp of AI technologies in their fields, he says.
But academic-driven AI initiatives present a problem for higher ed IT leaders, Hudson says: “The challenge for me as a CIO is, how do I give our faculty, researchers and students a playground for AI, but in a way that maintains the sovereignty of our core systems?”
Part of the challenge is that AI initiatives are rolling out much more quickly than traditional system-replacement cycles in higher education. “A university might spend two or three years selecting and implementing a new ERP, while an AI tool can be piloted in a matter of weeks,” Ménard says.
DISCOVER: Risk frameworks help institutions safely implement artificial intelligence.
Amid such rapid innovation, however, higher ed IT leaders must be very intentional about how they invest in AI. “We don’t want to go out and buy something that’s the latest and greatest, and then six months from now, it’s not what we want,” Hudson says.
Enabling AI With a Stable Digital Foundation
AI initiatives aren’t wholly distinct from core systems. The fact is that AI needs these systems — and needs them to run well.
AI highlights the necessity of a stable digital foundation, Ménard says. “If anything, AI is increasing the importance of having clean data, integrated systems and modern infrastructure.”
Despite all of the current attention on AI, “core systems haven’t become any less important,” he says.
While they pursue AI innovation, institutions need to continue investing in ERP, learning management systems, student information systems, customer relationship management and cybersecurity. “That’s because these systems form the foundation for everything else,” Ménard says.
READ MORE: Solid artificial intelligence requires solid infrastructure.
For instance, at KU, administrative staff members are exploring how they can use AI agents to complete repetitive tasks, such as managing email and calendars. “But agentic AI would need access to a whole host of databases and core systems,” such as CRM systems, Hudson says.
AI also can power insights into student success and retention. But to fully benefit from AI analytics, institutions need good data quality and system integration.
“In other words,” Ménard says, “the AI conversation often leads back to foundational work.”
Establishing Good Governance
The push for AI poses a challenge not only to systems but also to the people who manage them.
At KU, Hudson’s team regularly fields requests from faculty and researchers to purchase new AI tools. In addition to their everyday work, the IT team methodically assesses these requests, determining what each tool would do and what data it would access. The team reviews each AI tool’s security, privacy and compliance.
Hudson’s team also ensures that users are aware of the university’s data classification policy, which specifies the data that can and cannot go into AI tools, and that vendors agree to the school’s AI guidelines, such as not having their systems learn on university data.
UP NEXT: Is your data governance ready for artificial intelligence?
“Our constant push and pull is taking care of the existing systems while adding new ones,” Hudson says. “That puts a drain on our resources.”
To mitigate the issue, Hudson is considering adding more IT staff and upskilling current staff on AI-related skills. In addition, KU is forming a governance steering committee composed of senior leaders who will review larger AI projects and investments.
What Successful Balance Looks Like
Institutions that effectively balance AI innovation with core systems don’t treat AI as a separate initiative, Ménard says: “They’re focusing on building a strong foundation first.”
These institutions also start with targeted use cases, he adds. “They pilot, learn, establish governance and then expand.”
Success doesn’t mean the fastest, most extensive AI adoption, he notes. “The institutions that are likely to get the most value from AI over the next few years won’t be the ones adopting the most tools. They’ll be the ones that combine AI experimentation with strong operational discipline.”
