Skip to content

Lending platform

Credit decided, priced and disbursed in one system — and the difference is measurable.

90%

Faster than traditional underwriting

25%

Larger portfolio

70%

Lower cost to run

Business impact · 01

From application to funds released, with no human in the path.

I · Speed

90%

AGAINST TRADITIONAL UNDERWRITING AND DISBURSEMENT PROCESSES

The traditional path is drawn as a long line that winds and loops back on itself four times before funds are released. The platform path is drawn beneath it at the same scale: straight, and ending after the first tenth of it.

Traditional · four loops before funds are released. This platform · application, scored in-line, offer accepted, NIP transfer.

A bigger book from the customers you already bank.

II · Reach

25%

LARGER PORTFOLIO

Campaigns
pre-approved offers over email, SMS, USSD and WhatsApp
Self-service
BVN, date of birth and phone verification, two-stage OTP — no branch visit
Auto-approve
tier-based approval widens the addressable band
SME scorecard
a separate 15-rule scorecard and DSCR affordability, alongside retail

Credit policy is a database row, not a release cycle.

III · Cost

70%

COST SAVINGS

memo → ticket → code change → QA → release train → deploy

Not in the path

Tabs · Scoring rules · SME scoring rules · Scoring tiers · Loan product · Affordability · SME loan · SME affordability · Collections · Notifications
Score budget · Longevity · Income · Balance health · Behaviour · Loan history · Risk flags
Redrawn from the platform’s Policies screen. Values shown are illustrative.

Top features

Five pieces of machinery, drawn rather than described.

01

Income · obligations · headroom

Affordability

Income, obligations, and where affordability bites.

DSR · DSCR · per-tenure offers

02

Maker signed · checker empty

Maker-checker

Two different people, or the decision does not move.

CBN dual authorisation · same user rejected

03

Current · 1–30 · 31–60 · 61–90 · 90+

Restructure · write-off · GSI

Collections and recovery

Arrears age themselves and leave by three doors.

Daily sweep

04

Gini · KS · AUC · PSI

McNemar’s test

Model monitoring

A challenger scores the same book, silently.

Shadow scoring · champion/challenger

05

One unbroken record

Every stage of every loan, one record.

Immutable, PII-scrubbed audit trail

250+ API endpoints · 31 routers · 17 tables · 9 RBAC roles · 4,500+ automated tests · PostgreSQL 16 with read replica · Redis 7 · Keycloak 26 · Azure, GCP and AWS via Terraform · 50K concurrent, load tested

The screen a credit officer works in, redrawn as it is.

Redrawn from the platform’s customer screen. Identity masked; every value slot drawn empty.

Score factors · Decision · Compliance — BVN, NIN, AML, PEP, no default, consent.

Decide — per-rule score explanation, tiers A–E, DSR and DSCR affordability, maker-checker on every credit decision.

250+ API endpoints · 31 routers · 38,000+ lines across 130+ modules · 47 React components · 4,500+ automated tests (161 unit / 62 integration / 19 adversarial / 27 security) · 17 ORM tables · 9 RBAC roles · PostgreSQL 16 + read replica · Redis 7 · Keycloak 26 · Azure primary, GCP and AWS via Terraform · 50K concurrent, load tested (p95 <1s, error <0.1%) · NIBSS · CRC Bureau · FirstCentral · CAC · GSI · Paystack · Remita, TLS certificate pinning on every one

Cybersecurity

Fourteen layers, and no single human who can read a customer.

01Edgecloud WAF, DDoS, geo-blocking
02App WAFauto-ban at 10 violations
03TransportTLS 1.3, SPKI certificate pinning
04CSPscript-src self, frame-ancestors none
05AuthKeycloak 26, OIDC, PKCE, FIDO2/WebAuthn
06RBAC9 roles
07ApplicationPydantic validation, CSRF, DLP scanning
08DataAES-256-GCM, HMAC search indexes
09Auditimmutable, PII-scrubbed, transaction-scoped
10IAM5 roles, separation of duties
11MonitoringFalco eBPF, 18 kernel rules
12Deception40+ honeypot endpoints, canary tokens
13IncidentNDPA 72h, CBN 24h registers
14Supply chainSLSA Level 3, cosign signing

CBN · FCCPC DEON · NDPA 2023 · BOFIA 2020

Run it against your own book and see which figure moves first.

Tell us the portfolio, the core banking system and the constraint. We will run the scoring engine against your own data.