Churn signals before the customer leaves
A telecom operator learns about churn when the contract is already cancelled. We link billing, tickets and usage to see the risk in advance.
This is a demo scenario: how we would build the solution. Names and data are illustrative — it is not a specific client’s story.
Task
Find subscribers whose churn risk is rising and hand them to retention while they are still customers.
Constraints
- Billing and support data live in different systems with different identifiers.
- Personal data never leaves the company perimeter.
Solution
- 01
An identifier mapping for each subscriber across systems.
- 02
A weekly risk model on usage, payment and ticket features.
- 03
The high-risk list goes to the retention team’s CRM with the reason.
Architecture
- Billing → ETL · ID mapping
- Support desk → ETL · ID mapping
- Usage records → ETL · ID mapping
- ETL · ID mapping → Data Warehouse
- Data Warehouse → Churn risk model, Churn dashboard
- Churn risk model → Retention CRM
- Retention CRM
- Churn dashboard
What it demonstrates
Demonstrates: joining data from systems with different identifiers and turning a forecast into a concrete action list for a team.
Want an AI agent to prepare an offer for each at-risk subscriber?
AI agentsAgentZone