---Business Summary Report: Predictive Insights for Collections Strategy
#B
---
## 1. Summary of Predictive Insights
Our predictive model identified key drivers of customer delinquency and high-risk segments to inform targeted collections
efforts:
* **Key Predictors:**
* **High Credit Utilization Ratio:** Customers using more than 75% of their credit limits face a significantly higher risk of
delinquency.
* **Prior Missed Payments:** History of late or missed payments strongly correlates with future delinquency.
* **Low Monthly Income:** Customers in lower income brackets with high credit usage are more vulnerable to default.
* **High-Risk Customer Segments:**
* Young adults aged 25–35 with credit utilization >75%, monthly income below INR 40,000, and previous missed
payments.
* New customers (less than 1 year) exhibiting rapid credit limit growth combined with high utilization.
### Key Insights Summary Table
| Key Insight | Customer Segment | Influencing Variables | Potential Impact
|
| -------------------------------------------- | ------------------------------------------- | ------------------------------------- |
----------------------------------------------------------------------- |
| High credit utilization predicts delinquency | Customers aged 25–35 with >75% credit usage | Credit utilization ratio, age,
income | Enable proactive credit limit adjustments and alerts to reduce defaults |
| Prior missed payments increase risk | Customers with prior payment delinquencies | Payment history, delinquency
flags | Prioritize collections outreach and risk monitoring |
| Low income with high credit use raises risk | Lower-income customers using >70% credit | Monthly income, credit
utilization | Tailored financial counseling and support programs |
---
## 2. Recommendation Framework
### Restated Insight:
High credit utilization (above 75%) is a leading predictor of customer delinquency.
### Proposed Recommendation:
* **Specific:** Implement automated alerts and credit limit reviews for customers exceeding 75% credit utilization.
* **Measurable:** Target a 15% reduction in delinquency rates within this segment.
* **Actionable:** Use existing customer data to trigger alerts and provide personalized credit counseling.
* **Relevant:** Aligns with Geldium’s goal to reduce non-performing assets by proactively managing credit risk.
* **Time-bound:** Launch the intervention within 3 months and monitor outcomes quarterly.
### Justification and Business Rationale:
This focused approach addresses a top risk factor, enabling the Collections team to intervene early and help customers
manage credit more responsibly. Reducing high utilization reduces financial stress on customers, lowering delinquency and
improving portfolio health.
---
## 3. Ethical and Responsible AI Considerations
* **Fairness and Bias Risks:**
* **Income Bias:** The model may disproportionately flag lower-income customers as high-risk, potentially leading to
unfair credit restrictions.
* **Historical Bias:** Dependence on prior payment behavior risks reinforcing disadvantage for customers with past
hardships.
* **Mitigation Strategies:**
* Regular audits of model outcomes across income and demographic groups.
* Introduce human oversight for borderline cases to ensure fairness.
* Transparently communicate reasons for credit decisions to customers.
* **Explainability:**
Model predictions are based on clear, understandable factors like credit utilization and payment history. Collections teams
can explain risk scores in plain language, increasing transparency.
* **Responsible Financial Decision-Making:**
The recommendation supports responsible lending by encouraging manageable credit use, helping customers avoid over-
indebtedness.
* **Additional Ethical Principles:**
* Ensure data privacy and comply with regulations in handling customer information.
* Maintain accountability through regular model performance reviews.
---
**Prepared for:**
Geldium Head of Collections
Shreya jain
11 August, 2025
