Eneco churn prediction

par Eneco

AI Application

Faible confiance

Eneco's data and machine-learning staff predict customer churn with deep-learning survival models, which estimate when a customer is likely to leave rather than only whether. Eneco presented the work at PyData Amsterdam 2025 as a case study used to improve long-term planning and customer retention. The talk was given by the lead of data and machine learning in Eneco's Energy Trading teams and a senior data engineer, alongside a second case on a high-availability real-time trading data system built on Databricks. Eneco's December 2025 privacy statement says it uses profiling to analyse or predict preferences, interests and visitor behaviour, can use purchased data for this, and takes no automated decisions with legal or similarly significant effects based on profiling. The talk page does not state hosting or how retention staff act on the predictions.