Probing the hype between the pipes / Alence Poudel and Trevor Surface.

Abstract: With growing frequency, procurement requests from local and state governments are including phrases such as "predictive analytics" or "AI-enabled asset management." The assumption is that machine learning (ML) will improve outcomes because it has worked elsewhere. The city of Sugar Land, Texas, tested that assumption to determine if ML could improve how the city prioritizes pipe repairs and upgrades using only the kinds of data most utilities already have, such as work orders, basic inventory, and consequence ratings. Sugar Land's experience suggests that traditional risk scores remain reliable under typical municipal data constraints. When data maturity increases, the equation may change.