On December 4, 2025, the UWEBC hosted a cross‑functional Data & Analytics and AI Interest Group session focused on one of the most pressing challenges facing organizations today: how to innovate boldly with AI while maintaining responsible governance. Led by UWEBC Director Doug Barton, this highly attended virtual event brought together nearly 150 participants from close to 50 member companies for a morning of expert insights, candid practitioner stories, and interactive discussion.
The session opened with remarks from Doug, who framed the day around the growing tension organizations face: innovation versus oversight, speed versus safety, performance versus transparency, and AI autonomy versus human judgment. He emphasized that while AI is accelerating rapidly, most companies are still early in their journey, making this a critical moment to learn from peers, share dilemmas, and build intentional strategies for the future.
The first spotlight presentation featured Cody Baldwin, Director of the Master’s in Business Analytics program at the Wisconsin School of Business. Cody shared five common mistakes that threaten AI initiatives, including falling in love with technology instead of business problems, neglecting the “soft stuff” like expectations and prioritization, automating broken processes, relying on untrusted sources, and assuming everyone else is further ahead than they really are. He reminded participants that most organizations are still “crawling” with AI, not running, and encouraged leaders to focus on fundamentals: domain expertise, business context, and thoughtful problem framing.
“We think everyone is running with AI when most people are still crawling.” — Cody Baldwin
Following Cody, participants heard from Josh Murray, Director of AI, Innovation, and Emerging Technology at Alliant Energy. Josh walked through Alliant’s multi‑year journey—from early analytics teams to a unified data science function to today’s enterprise AI strategy. He shared real examples of dilemmas his team faces, such as whether to build AI solutions now or wait for vendor roadmaps to mature, how to democratize generative AI while maintaining guardrails, and how to help employees understand the limitations of AI tools. Josh also highlighted Alliant’s shift from dashboards to insights, their use of Snowflake to scale data access, and the importance of pairing experimentation with governance.
“One of the biggest challenges is balancing governance with innovation. These tools are powerful, but they will lie to you with confidence.” — Josh Murray
The next spotlight came from Nathan Lasnoski, CTO at Neudesic, an IBM Company. Nathan shared insights from Microsoft Ignite and demonstrated how rapidly enterprise AI capabilities are evolving. He introduced the concept of “digital workers” and showed how organizations are already using AI agents to automate customer service, claims adjudication, document processing, and even software development. Nathan emphasized that the biggest differentiator is not the technology itself, but the intentionality behind selecting the right problems and designing guardrails that enable AI to perform real work safely.
“What if 50% of the work in your organization could be performed by digital workers?” — Nathan Lasnoski
Participants then heard from Mingju Sun, Vice President of Data and AI at American Family Insurance. Mingju shared how AmFam is building an enterprise AI platform that supports both task‑based and orchestration‑based AI agents. She emphasized the importance of strong data foundations, clear workflows, and thoughtful architecture to support monitoring, observability, and responsible deployment. Mingju also reminded participants that not every problem requires AI—sometimes simple data and coding solutions are more effective.
“Not everything needs to be AI‑fied. If you can solve it with data and simple coding, you should.” — Mingju Sun
The final spotlight presentation featured Yonatan Mintz, Assistant Professor of Industrial & Systems Engineering at UW–Madison. Yonatan shared research from the Mintz Lab on behavioral analytics and just‑in‑time interventions, highlighting a recent collaboration with American Family Insurance. His team used telematics data and reinforcement learning to send personalized text nudges that helped drivers reduce hard‑braking incidents. He walked participants through the trial design, the algorithmic approach, and the broader implications for personalized, data‑driven interventions across industries.
The event concluded with breakout discussions, where participants connected in small groups to explore their own dilemmas around AI governance, experimentation, data quality, and organizational readiness. These conversations allowed members to learn from one another, compare approaches, and gather practical ideas to bring back to their teams.
“I’m going to work with my team to identify what they struggle with daily and eliminate non‑value‑add tasks.” — Event Participant