“"It doesn't fix the problem. It does the troubleshooting for you. Proactive instead of reactive."
Kenneth BurdineAssociate VP and Deputy CIO @ Kentucky Community and Technical College System
At a recent conference, most CIOs had the same answer when asked about AI: wait and see. Let someone else be the guinea pig. Too many things could go wrong. That posture is understandable given the hype cycle, but it also means the organizations waiting it out aren't learning anything about where AI actually works. At one community college system in Kentucky, the approach is different: pick the operational problems that eat the most time, find the AI-enabled tools that address them inside trusted vendor products, and measure whether they actually shortened the gap between investment and usable capability.
Kenneth Burdine is Associate Vice President and Deputy CIO at KCTCS, the Kentucky Community and Technical College System, which serves more than 100,000 students across 16 colleges statewide. He oversees enterprise infrastructure, networking, and IT operations for the system, including more than 3,000 wireless access points. He doesn't describe himself as an AI advocate. He values common sense, wants his teams to prevent problems instead of firefighting them, and needs new staff to be productive in weeks rather than years.
"I didn't need AI to do the work for me. I needed it to give me an architectural overview, and I needed the capability to troubleshoot common services," Burdine said. "It doesn't fix the problem. It does the troubleshooting for you. Proactive instead of reactive."
KCTCS had recently deployed Juniper's AI-enabled Mist platform for wireless and wired networking, replacing a Cisco infrastructure that had been in place for years. Rather than brand loyalty, the decision came down to what the product could do for an organization that couldn't afford to be reactive with limited staff spread across 16 campuses.
Training in days: Burdine recounted a story from the evaluation process. A help desk employee with no networking background received a 15-minute tutorial on the Marvis AI interface. Two hours later, the person was already identifying issues and finding their way around the system architecture. "I've been told that normally it takes six months to a year through the Cisco Academy," Burdine said. At KCTCS, where staff frequently leave for other positions after being trained, that acceleration changed the ROI math. "Normally it takes us a couple of years to get someone up to speed. Now I can get people there in 30 to 90 days. The return on investment is that I got them for another 15 months before they go looking for a better job."
Proactive troubleshooting at scale: With 100,000 students who don't report wireless issues until they're genuinely frustrated, and don't describe them well when they do, KCTCS needed infrastructure that could identify problems before tickets arrived. The new platform gives visibility down to the application layer, detecting bad cables, connection failures, and service degradation across all 16 colleges in real time. "Is it easier to patch the boat before it gets a hole in it, or sit there bailing out water?" Burdine said. "We can see these issues before they start affecting students."
Cross-college support: Because all 16 colleges now run the same standardized environment, IT directors can monitor each other's infrastructure during major deployments. If one college needs to put staff on the ground for a hardware rollout, a peer at another campus can watch the wireless and wired services remotely. "We didn't have that capability before because every environment was different. Now I can literally click and see everything," he said.
The deployment itself was the most concrete proof point. KCTCS's previous route/switch deployment on the old platform had taken roughly two and a half years. The Juniper rollout took less than six months. But Burdine credited the speed to preparation as much as the product. Before a single device was installed, his team standardized VLANs system-wide, identified integration requirements across all campuses, and Burdine personally spoke with each of the 16 IT directors to understand their environments and concerns. "Communication is what makes AI implementation successful. Getting people prepared and getting input from the people who have to support it. That's what made it work," he said.
Burdine was direct about what AI should and shouldn't do. Stick with trusted vendors building AI into products using decades of operational data, not bolting novelty features onto untested platforms. Don't introduce something staff can't govern. And don't lump ROI metrics together. Training time, deployment speed, proactive issue detection, and cross-campus collaboration each have their own return and need to be measured separately to tell a credible story.
"Remember when pneumatic hammers came into the construction industry? Guys said it was going to take their jobs," Burdine said. "It didn't. It made them more productive. They could build houses faster. That's all AI is. A force multiplier. If it doesn't make common sense for your environment, don't do it."
