She Used Data to Add More Love to the World. It Closed the Achievement Gap.

Dr. Bernadette Howlett has spent her career at the intersection most people don't think about: where data science, machine learning, and human-centered education meet.  As Operations Analytics and Experiential Product Director at Western Governors University, she leads a team using predictive modeling and decision intelligence to help students finish the credentials they started faster, with better outcomes, and with less debt.   Her mission has a name: Stop the drop with debt Because a student who leaves without a credential but with loan payments is the worst possible outcome and it's preventable.   In this conversation, Dr. B breaks down how WGU deploys a 78-feature predictive mode to:  - Flag student momentum - Surface personalized faculty outreach through a language model called FOME (Faculty Outreach Message Engine) - Uses social norming principles to nudge students toward action The faculty always reviews and can modify before anything goes out.  The human is always in the loop.   She also shares the lesson that almost derailed the whole system: well-intentioned SLA dashboards that faculty loved in theory and resented in practice because managers turned them into performance metrics.  The fix? Remove the dashboard entirely when they launched decision intelligence.   Other highlights from the conversation: - How WGU reduced accreditation reporting from weeks to hours without removing human verification - The employer who stood up in a meeting and said flatly, "We don't trust you" and why Dr. B thinks he was right - Her decision-making persona breakdown: 50% values, 30% strategy, 10% contrarian, 10% evidence-based - Why she calls evidence-based practice "the vessels that feed the heart"   If you work in higher ed, workforce development, or anywhere humans are trying to use data to help other humans, this one's for you. cmodel.io