Predicts which customers are most likely to leave, explains the strongest churn drivers, and produces an exportable retention queue for CRM or customer-success teams.
The examples below show where this project can be used, what decision it supports, and how a user can test the scenario in the sandbox.
Where it fits: Use tenure, contract, usage, support history and payment behaviour to identify customers likely to disconnect or downgrade.
Decision supported: Prioritise retention campaigns, tariff review, service recovery and targeted offers.
How to test it: Load the sample data, train the model, score current customers, then sort by high-risk customers and top drivers.
Where it fits: Score customers with declining transactions, fewer deposits or reduced app activity before they become inactive.
Decision supported: Trigger personalised engagement through relationship managers, app nudges or product bundles.
How to test it: Upload a customer activity CSV and map churn/attrition as the target.
Where it fits: Detect policyholders likely to lapse based on premium history, claims experience, product type and engagement signals.
Decision supported: Protect renewal revenue and focus retention outreach on high lifetime-value customers.
How to test it: Use the exclusion checkbox to remove ID/leakage fields, then inspect retention opportunity KPIs.
Where it fits: Combine product usage, support tickets, billing plan and customer health metrics to flag accounts at risk.
Decision supported: Prioritise customer success interventions and renewal planning.
How to test it: Use the assistant to ask which risk drivers should be handled first.
Where it fits: Apply churn logic to programmes where beneficiaries, partners or trainees stop engaging across a multi-stage journey.
Decision supported: Improve continuity, completion rates and programme value for donors.
How to test it: Use the scored output as a retention queue and export action recommendations.
The examples above were revised against current industry and research references on how this class of analytics solution is used in practice.