You Can Adopt AI Without Losing Control

AI in credit unions can feel like a risk until you see it as a workflow you can test, measure, and control. We’re joined by Eric Burgess, Director of Home Loans at First Commerce Credit Union, to talk about what it really takes to get comfortable being uncomfortable and start using artificial intelligence in mortgage lending without handing over the keys. Eric shares how he delved deeply into AI training and why he helps others learn through the AI Learning Lab community.

We get practical fast: prompt engineering as the starting point, why “human in the loop” is non-negotiable, and how leaders should think about data security when employees use public LLM tools like ChatGPT. Eric breaks down a crawl-walk-run approach for credit union AI adoption, including a simple but powerful pilot: running vendor contracts through an AI agent in seconds, then comparing the output to attorney redlines that can take weeks.

From there, we move into agents and automation using tools like Microsoft Copilot Studio, including a mortgage quality-control agent that flags risks and dramatically reduces QC time while leaving final decisions to people. We also explore what “advanced” can look like, from AI-assisted underwriting that boosts underwriter productivity to predictive analytics dashboards that improve planning and forecasting.

If you’re looking for real AI use cases in credit unions, mortgage operations, and lending productivity, this conversation is a strong place to start.

Subscribe, share the episode with a colleague, and leave a review to help more credit union leaders find it.

Sponsored by Xactus