The MSME Financial Health Card

Sees the invisible borrower — in both directions.

Six crore MSMEs run real businesses that credit bureaus cannot see; others look clean on paper while their cash flow quietly burns. SHROFF reads the consented ledger — bank, GST, EPFO — scores it with an explainable model, and screens every identifier against the public negative registries. One card. Both failure modes.

6.3 Cr
registered MSMEs in India
₹25L Cr
estimated formal credit gap
24 mo
consented cash-flow window
< 60 s
score, reasons & sanction
01

Cash-flow truth, consented

24 months of bank inflows, GST filings and EPFO payrolls fetched over Account Aggregator rails under a purpose-bound ReBIT consent. The borrower's real ledger — not a credit bureau's silence.

02

Deterministic score, signed reasons

Four monotonic gradient-boosted sub-models — cash flow, growth, stability, compliance — calibrated to a 12-month PD, each decision carrying TreeSHAP reason codes. Nothing generative ever computes a number.

03

Registry screening & the entity graph

Every identifier walked through MCA struck-off, GST non-genuine, wilful-defaulter and DIN registries — then along the promoter graph, where phoenix operators hide behind fresh CINs at old addresses.

Direction one — the deserving invisible

A kirana store with no credit history but two years of clean, growing, GST-consistent cash flow deserves working capital. The onboarding head prices it: score, band, sanction amount under the 20% working-capital norm, tenure — with every reason signed.

Direction two — the clean-on-paper risk

Declared turnover 40% above bank receipts. Bounces stacking up. A promoter whose last company was struck off in May and reborn in June at the same address. The early-warning head and the entity graph catch what a bureau score never will.