United KingdomPublic sector2025 – ongoingDepartment for Work and Pensions (Child Maintenance Service)
Predictive analytics for payment compliance
Child Maintenance Service model flags payers likely to miss a payment, scoring about 100,000 cases a month
Problem
Caseworkers at the Child Maintenance Service want to step in before a compliant paying parent misses a payment, which can start a longer breakdown.
Approach
An XGBoost model with explanations flags compliant cases at risk of missing a payment. It was rolled out to a wider group of caseworkers in March 2025, and more than 40 caseworkers now use it. A new model must reach a precision of at least 0.62 to be accepted.
Results
Takeaway
The model flags cases where staff can intervene before a missed payment. The service reported model accuracy and precision but no measured effect on continued payments.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Production and qualityFinance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The Child Maintenance Service described the model, its training data and the checks carried out before use. It reported accuracy, precision, monthly volume and the number of caseworkers using it.
Full details are in the original sources.
