Collecting intelligently: fewer calls, more cheques
Collections on delinquent mortgages can often get exasperating, requiring multiple follow-up calls with customers. But they can also get exhilarating – when done right. That’s what James Gibson, Head of Collections Strategy at a top 10 US bank discovered when he brought in an Infosys BPM team to streamline his collections processes. This case details how the team, deploying classical AI models to segment customers based on historical payment behaviour, and automating many of the earlier indiscriminate follow-up calls, delivered close to $1.3 Mn in automated collections.
Approach summary:
- Evaluation of historical customer payments
- Application of classical AI algorithms
- Segmentation of customer accounts
- Automation of differentiated outreach calls
- Explainability with rule-based logic
- Phased implementation and continuous monitoring
Key benefits:
~$1.3M Mn in automated collections 
~13% reduction in agent-driven calls 
$558K in recoveries through redeployment 
