Composite Credit Scoring
One Score. Every Data Source. Fully Explainable.
The AgenticLending composite credit score synthesises CIBIL bureau data, GST filing history, ITR income declarations, and bank statement cash flows into a single, risk-calibrated credit assessment — purpose-built for Indian commercial lending, with a full explanation your credit committee can stand behind.
Data Inputs
Four Sources. One Unified Credit View.
The composite score is built from verified data across every major Indian credit data source — not just a CIBIL score repackaged. Each source contributes independently weighted signals to the final assessment.
CIBIL Bureau→
Credit score, DPD history, existing obligations, suit-filed accounts, and enquiry patterns.
GST Filing Analysis→
Declared turnover, filing compliance, ITC ratios, and multi-period revenue trends.
ITR Income Analysis→
Verified income, DSCR-ready P&L, balance sheet data, and cross-source income triangulation.
Bank Statement Analysis→
Actual cash flows, average balances, anomaly flags, and cross-validated bank-implied turnover.
How Scoring Works
Deterministic. Explainable. Calibrated for Indian Credit.
The composite engine uses a deterministic scoring model — not a black-box ML algorithm. Every score component is traceable to a specific data point, making the output defensible for RBI inspections and internal audit.
Weighted Multi-Source Model
Each data source — bureau, GST, ITR, bank — contributes a weighted score component. Weights are configurable per product type (working capital, term loan, LAP) and borrower segment.
Cross-Validation Penalties
Material inconsistencies between data sources — e.g., bank-implied revenue significantly above GST-declared — automatically reduce the composite score with a flagged explanation.
Behavioural Trend Scoring
Trend components reward consistent GST filing, stable bank cash flows, and improving credit bureau scores — penalising recent deterioration signals even if absolute levels are acceptable.
Product-Specific Calibration
Score weights and cut-offs are configurable per lending product. Working capital underwriting weights bank cash flows more heavily; LAP underwriting weights asset cover and ITR income.
Director & Promoter Overlay
Individual bureau and income data for directors and promoters is incorporated as a risk overlay — surfacing cases where entity credit looks acceptable but promoter stress is elevated.
Score Explanation Layer
Every composite score comes with a plain-language explanation — the top factors driving the score, the key risk flags, and the data sources behind each component.
Key Outputs
Credit Committee-Ready Assessment
Composite Score & Band
Single numerical score on a standardised scale with risk band classification (High / Medium-High / Medium / Low) and confidence indicator based on data availability.
Score Decomposition
Component-level breakdown showing exactly how much each data source contributed to the final score — with the strongest positive and negative drivers highlighted.
Risk Flag Register
Consolidated list of all risk signals surfaced across data sources — cross-source variances, compliance gaps, bureau alerts, and behavioural anomalies — ranked by severity.
Credit Narrative
AI-generated credit assessment paragraph summarising the borrower profile, key strengths, principal risks, and underwriting considerations — ready for credit memo insertion.
Compliance & Audit
Every Decision. Documented and Defensible.
AgenticLending maintains an immutable audit trail for every composite score — the input data, scoring model version, weight configuration, and output score are all stored and queryable. Your compliance team can reconstruct any credit decision for any inspection, at any time.
RBI Fair Practice Code
Score explanations satisfy RBI guidelines on communicating credit decisions to applicants — plain language, no black-box outcomes.
Model Version Control
Every score stores the exact model version and weight configuration used — so historical decisions remain reproducible after model updates.
Override Logging
Credit officer overrides are logged with the overriding user, rationale, and authority level — maintaining accountability at every step.
Ready to Transform Your Credit Decisioning?
The platform is live. Try it now or book a 30-minute guided walkthrough.