Skip to main content

Case studies

Preventive complaint handling

For companies in the energy services sector, strict national regulation makes it essential to continuously monitor the dialogue with customers. With the pilot solution, EON can predict cases that cases that threaten consumer protection action.

It is the largest non-state-owned player in the domestic energy services sector.


The National Consumer Protection Authority imposes fines the company in a number of service-related complaint cases. During investigating the background of the relevant cases, E.ON finds that in many cases that there was a prior dialogue between the two parties. How can a company with a professional customer service department identify cases that could lead to a consumer protection complaint by analysing the communication and avoid a fine?


E.ON has developed a pilot solution which was able to predict up to 70% of consumer complaint cases and reduce the amount of fines imposed on the company.


By segmenting customers, predictive models based on communication data specific to each group are used to predict which cases require special attention.

  1. Preparation
    • developing customer segments:
      • Universal customers - general private users
      • Competitive customers - enterprises
    • Analysing 1,5- 2 years of written or voice-based customer communication data
  2. Implementation
    • Building predictive models from structured data
    • Evaluation: comparison of 2 months of complaint cases with the model's predicted complaint threat - 70% accuracy
    • analysis can take place monthly