Regional internet provider
Zum Telecom
The provider's ERP and the support system cross-checked — a model that points to who is likely to cancel.
The context
Zum Telecom is a regional internet provider. Like every provider, it lives with cancellation — churn — as the main threat to revenue. The signs that a customer is about to leave do exist, but they sit scattered between the system that handles the contract and the system that logs support.
What wasn't connected
On one side, the provider's ERP: contracts, plans, invoices, each customer's history. On the other, the support system: every ticket, every complaint, every conversation. Two worlds that didn't talk. No one could look at a customer and see, at once, both the contract and the friction history — and it's in the crossing of the two that the warning of departure lives.
What DUBAI connected
DUBAI crossed the two systems into a single history, with millions of records aligned per customer. On top of that base, a predictive model began scoring, customer by customer, the risk of cancellation — reading together what the contract shows and what support reveals. What was scattered became a single signal.
Signature moment · The warning before cancellation
The turning point is simple to tell: the system points to who is likely to cancel before the customer cancels. Instead of discovering the loss once the contract is already closed, the operation gets the name of the at-risk customer while there's still time to act. Retention stops being a reaction and becomes anticipation. Today, the system runs in production.
Results
- 2 systems
- the ERP and support cross-checked per customer
- millions
- of records in a single history
- per customer
- cancellation risk scored, one by one