Lapse prediction, retention offers, churn prevention and win-back on one side; cross-sell, upsell and lead scoring on the other. Scores are computed from each record's features; the next best action is reasoned by the model over those scores — retain, cross-sell, win-back, fix service, or leave alone.
Active policies with the highest computed lapse propensity. Each row can ask the model for the best next action, given both its lapse and cross-sell scores — with a script the agent or LEO would actually use.
Live rows: 300 policy records read from demo_bfsi.insurance_policy in PostgreSQL at request time. These are genuine records with genuine structure, but they are synthetic demo data, not Generali Central's book — status distributions are near-uniform, and person, provider, branch and payer names are localised to Indian display names at read time. Scores demonstrate the scoring mechanism, not predictive performance. The reconciled ₹5,548 Cr figures elsewhere in this cockpit come from the governed SQLite book.
Single-product customers with income headroom, current KYC and clean claims score highest; the engine also names the product gap to offer.
| Customer | Holds | Income | Score | Offer | Why |
|---|---|---|---|---|---|
| Meena Pandey | MOTOR | — | 0.68High | HEALTH | Single-product customer +0.18KYC verified +0.10Low risk +0.10 |
| Dinesh Bose | HEALTH | ₹2 L | 0.58Medium | MOTOR | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Kartik Shah | HEALTH | ₹1 L | 0.58Medium | MOTOR | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Vikram Mishra | LIFE | ₹3 L | 0.58Medium | HEALTH | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Rekha Naidu | TRAVEL | ₹1 L | 0.58Medium | HEALTH | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Kavita Banerjee | MOTOR | — | 0.58Medium | HEALTH | Single-product customer +0.18Low risk +0.10 |
| Deepa Chauhan | HOME | ₹1 L | 0.58Medium | HEALTH | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Rekha Patil | HOME | ₹3 L | 0.58Medium | HEALTH | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Tarun Tandon | HEALTH | ₹3 L | 0.58Medium | MOTOR | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Ramesh Menon | LIFE | ₹3 L | 0.58Medium | HEALTH | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
| Sachin Naidu | TRAVEL | — | 0.58Medium | HEALTH | Single-product customer +0.18KYC verified +0.10 |
| Ankit Chowdhury | HEALTH | ₹0 L | 0.58Medium | MOTOR | Single-product customer +0.18Low income -0.10KYC verified +0.10 |
Prospect quality from income, KYC readiness, risk rating and the loss ratio of the branch that would write the business. No separate lead table exists in the source, so the customer base is scored as if each were a fresh lead.