An IDB-funded pilot equipped Peru's public utility EPS Grau with Idrica's GoAigua AI leak- and fraud-detection software in Piura's Sector IV, flagging over 20 loss events and cutting unbilled water by 15% within about a year.
15%
Non-revenue water reduction (~2021-2022 pilot)
>20
Loss events flagged (pilot duration)
Details
Maturity
Pilot
Promoter
EPS Grau S.A.
Period
2021–2022
Keywords
water utilities, public infrastructure, leak detection
Context
Peru's public water utilities lose a large share of treated water before it reaches customers; in Piura, state-owned utility EPS Grau sought to test whether AI-based analytics could reduce non-revenue water in a defined network sector.
Activities
Funded by the Inter-American Development Bank alongside Peru's Ministry of Housing, Construction and Sanitation, EPS Grau ran a roughly one-year pilot (completed 2022) deploying three modules of Spanish vendor Idrica's GoAigua platform in Sector IV: FlowSens (leak/fraud flagging via ML on sensor data), Meter Insights (smart-meter analytics) and Water Twin (SCADA digital-twin layer).
Results
Over the pilot, the system flagged more than 20 distinct loss events — undeclared leaks and abnormal consumption patterns — and EPS Grau recorded a 15% reduction in unbilled (non-revenue) water during the monitoring and operating phase.
Conclusions
The pilot covered a single network sector rather than EPS Grau's full service area, and no independent post-pilot audit of the sustained reduction has been published.
Rwanda's state water utility WASAC piloted smart meters in Kigali and worked with researchers applying machine-learning predictive modelling to nearly …