Close

New Research from CDW Explores AI and Cybersecurity

Learn how AI is helping IT teams manage risk and improve resilience.

Nov 19 2026
Artificial Intelligence

University Hackathon Aims To Foster AI Skills Among IT Staff

The two-day event at the University of Chicago targeted IT team members who had no software development experience.

At the University of Chicago, artificial intelligence tools, policies and training opportunities have been rolling out across campus. But IT leaders noticed implementation wasn’t keeping up. 

“AI was deployed, but adoption did not really follow,” says Valerie Archambeau, director of communication technology services at UChicago. 

Instead of producing another round of training, she and a group of IT colleagues devised a hackathon geared toward IT Services staff. Ultimately, 81 staff members formed 17 teams, working in person and remotely, and built more than a dozen working prototypes with Claude Code at the event. Here’s how the team made it happen.

Event Aims To Bolster AI Adoption Across Campus

UChicago’s AI rollout began in 2024 with a variety of AI tools, including Zoom AI Companion, Microsoft Copilot and an in-house ChatGPT instance called Phoenix AI. Despite a dedicated website with resources and training opportunities, users were not adopting AI as readily as leaders hoped. The team had to look for creative ways to get the community on board. 

Click the banner below to find tools to enable AI innovation on your campus.

 

“One of the things our leadership was looking for was getting people to experiment and try using code in a safe, fun environment,” says Allison Kallo, manager of learning design at UChicago. 

Held in March 2026, the hackathon started as a way to get developers comfortable with AI coding tools. But soon the team realized that the event needed to extend beyond those who already had coding experience. 

“Originally, the thought was, ‘Let's show our developers what they can do with this,’” Archambeau says. “By the very next meeting, it was like, why are we limiting this to developers? The whole point of this is that you don't have to be a developer.” That’s when the scope of the event extended beyond central IT to other team members across ITS. 

The goal was not as basic as teaching participants how to use AI for coding.

“The focus was on the problems that they’re seeing that this technology can address,” Archambeau says. “This is about how people use code to solve problems, simplify processes or make things possible that would be time consuming if done without the tool that provides that connection.”

DISCOVER: Vibe coding helps higher ed institutions solve problems.

AI Hackathon Welcomed IT Staff With No Coding Experience

Participation in the event was voluntary, so attendees showed up with curiosity about what the technology could do for them and how they could use it effectively, Kallo says. 

The ITS team offered guidance ahead of the event, so participants were prepared. Information sessions allowed the team to demonstrate AI’s capabilities and learn about the problems they wanted to solve. Representatives from Amazon Web Services were also onsite to answer participants’ questions, and attendees received a temporary AI enterprise license to use during the event.

“We also had office hours, where people could talk to us one on one,” says Astrid Fingerhut, identity and access management solutions and operations manager for UChicago. “So, you weren’t going to show up and be the only one who didn’t understand it. There was an opportunity to touch base with somebody who could help you and make you feel like, ‘Yeah, I can do this.’”

Tom Kompare, manager of application development at UChicago, says his team members already had some experience with the tool, so they knew how AI-enabled coding could change workflows and solve problems. He created a short guide to help participants get familiar with the software development process — gather requirements, design, build, test, present — and the AI tool itself.

Click the banner below to subscribe to our weekly newsletter.

 

Two-Day Event Resulted in Working Prototypes

Getting the right combination of people on each team took some engineering of its own. Staff submitted project ideas and then signed up by self-reporting their skills and experience, from never having touched an AI tool to using it daily. The hackathon team split participants into teams ahead of the event.

"We made sure that when we put people in groups that there was a different level of AI experience and professional skills, and hopefully the same interest in their team’s chosen idea,” Kompare says. To speed up the first pass at creating teams, he wrote a prompt for the AI tool to suggest assignments, then reviewed the results himself.

On the first day of the hackathon, participants, equipped with the guidelines they received in advance, got started on designing and building their tools with their teams. Mentors, including Kompare, UChicago IT Manager Arjun Lagisetty and several AWS staff, were on hand in case teams got stuck. By the end of the first day, Kompare says, many teams had worked through requirements and design and were nearly finished building.

On day two, teams had the morning to finish and test their projects and put together a three-slide presentation, which they delivered in the afternoon.

EXPLORE: Institutions are adopting a human-centered approach to IT.

The projects the teams built were mostly aimed at everyday workplace tasks, and a few of them are still in use.

The HR team built a tool that helps HR and hiring managers write consistent, compliant job descriptions faster. It classifies a role, matches it to the official job catalog and job code, and drafts descriptions from standard templates and approved language. The tool replaces writing each description from scratch and gives the user a draft to edit before it goes out. Kallo says the prototype became the base of a tool the department now uses.

Another team built a synthetic data generator that allows a developer to describe what they’re building and what data it needs, then produces data sets to match — deliberately messy, with edge cases included. And in another project, a language-learning Chrome extension swaps selected words on any web page into a foreign language, so users can guess the meaning from context as they read. 

In addition to the hackathon, the university has started a monthly AI user group, open to anyone on campus, where people demo what they’ve made. Archambeau says colleagues regularly use the chat to ask to see the prompts behind a demo so they can try something similar.

Tom Kompare
We made sure that when we put people in groups that there was a different level of AI experience and professional skills, and hopefully the same interest in their team’s chosen idea.”

Tom Kompare Manager of Application Development, UChicago

Guidelines and Guardrails Helped Direct Participants

For an event where dozens of nondevelopers would be typing prompts into an AI tool, the data rules were simple and set early. Participants did not have access to production systems or test environments. Anyone working with structured data had to sanitize it first; no student records, no health data and no personal information of any kind could be used.

“That was a directive: Bring your own data, but you have to fake it,” Kompare says.

The institution’s existing agreements with its AI vendors helped, Archambeau says. University data is not allowed to leave the institution’s environment or be used to train the vendor’s tools. She advises other institutions to meet with their security teams early because policy maturity varies widely.

The organizers also brought in the head of IT’s accessibility group to talk to participants about building with accessibility in mind and gave them a prompt to add to any request, asking the tool to meet Web Content Accessibility Guidelines (WCAG) 2.1 AA standards, with sufficient color contrast, alt text and keyboard navigation.

LEARN MORE: The CDW AI Developer Challenge allowed participants to create their own agents.

UChicago Team Offers Lessons and Resources for Other Institutions

The team learned several lessons throughout this process, which they plan to implement in the event’s next iteration early next year. 

While the inaugural event included a virtual element, with some teams connecting online, attendees found it was harder to engage in a virtual space. This option will be eliminated in the next event, Fingerhut says. Additionally, teams will be smaller, and the participants will have more time to demonstrate the tools they created on the second day.

The university’s AI environment and policies have evolved since the event, with employees having access to enterprise accounts and a better understanding of security parameters.

“We as an institution have more fleshed out security policies about how to use AI,” Kompare says. “We'll have better guidance than, ‘Don't use anything.’ We'll have policies to point people to.”

Building confidence in participants is key to empowering them to feel ready to experiment and create with AI. 

“I think too much handholding makes people afraid,” Kallo says. “I really think a little bit of cheerleading is good too. We did a pretty good job of being like, ‘Try it. You're not going to break it. And if you do, then you win.’ Having that kind of attitude was really important. That made people feel safe to explore.” 

To help other institutions replicate this event, UChicago team has created a set of resources that outline their process.

“You can see exactly what we did. You can use whatever tools are useful for you,” Fingerhut says. “This is what we did. Here's how you can do it. Here's how you can take what we've done and change it to suit your own environment and use it yourself.”

UP NEXT: Here’s how to approach licensing and budgeting for AI in classrooms.

skynesher/Getty Images