D
DurujScore
AI Credit Intelligence
Request demo
← Back to DurujScore
For lenders

AI Credit Intelligence for Banks, MFIs, SACCOs, and Digital Lenders.

Make faster, more explainable lending decisions using verified borrower signals, permission-based readiness reviews, and lender-controlled credit policies. Underwriting, validation, and outcomes in one closed-loop platform.

Request Demo →View API Platform
Lender platform

From application to performance feedback

UW
AI Underwriting

Score, probability of default, recommendations, reason codes, and human-review flags.

PL
Policy Decision Layer

Configure product rules, thresholds, approval paths, and versioned credit policy.

VA
Validation

Track calibration, score bands, cohort performance, and drift before models become stale.

OF
Outcomes Feedback

Ingest repayment outcomes so observed performance can validate future decisions.

GV
Governance

Audit trail, maker-checker controls, compliance evidence, and role-aware access.

API
API Platform

Connect by REST API, webhooks, CSV/SFTP, sandbox, or branch-ready console workflows.

Evaluation flow
  1. 1Review the underwriting workflowSee how borrower data, consent, identity, signals, score, and policy combine.
  2. 2Validate the platform on your portfolioCompare predicted risk against observed outcomes and portfolio segments.
  3. 3Deploy with governanceKeep lender judgment, policy ownership, and auditability at the center of every decision.
The closed loop

Underwrite → Validate → Ingest outcomes → Improve.

Most credit vendors stop at the score. DurujScore closes the loop so every repayment outcome informs the next decision and gives your risk committee defensible evidence the model works.

Underwriting

AI-recommended decisions, explainable end to end.

Every score comes with a default probability, recommended limit, recommended tenor, top reasons, fraud and human-review flags. The lender controls policy; the engine surfaces an auditable recommendation.

  • ✓Score 300–850 with risk tier and confidence
  • ✓PD%, recommended limit and tenor (policy hint, not offer)
  • ✓Top reasons array — never a black-box number
  • ✓Fraud warning + human review flag for the queue
Validation

Prove the model works on your portfolio.

AUC, KS, calibration by score band, drift detection — the metrics your risk committee needs before a policy change goes live. Run a backtest against your historical book or your pilot cohort.

  • ✓AUC with strength verdict (Strong / Acceptable / Weak)
  • ✓KS statistic + calibration table per score band
  • ✓Sample-size caveat when n is too small to trust
  • ✓Per-band gap between predicted PD and observed default rate
Outcomes

Close the loop with realised repayment data.

Ingest outcomes via CSV, SFTP or API. Late vs default vs written-off are distinguished; every row flows back into the validation report. The model improves with each pilot cohort.

  • ✓Bulk CSV upload with per-row validation errors
  • ✓Status enum: current / paid_off / late / default / restructured / written_off
  • ✓Audit-grade idempotency: re-uploading is safe
  • ✓Outcomes join back to the originating decision for replay
By institution type

Built for the way your institution actually lends.

The same platform, configured differently per lender model. Policy, signals and workflow adapt to the borrower segment you actually serve.

Microfinance institutions

Agricultural, group lending, branch workflows.

  • ·Group-lending repayment discipline as a signal
  • ·Branch-officer review queue with reason codes
  • ·Outcome ingestion via CSV or SFTP
SACCOs

Member savings, guarantor chains, cooperative trust.

  • ·Member contribution history as a positive signal
  • ·Share-capital backing as collateral context
  • ·Guarantor chain visibility within the decision
Fintech lenders

Wallet-first, API-native, fast-decision flows.

  • ·Sub-second scoring via REST API
  • ·Mobile-money signals as primary input
  • ·Webhooks for score and decision events
Banks

Salary, SME, merchant, diaspora-backed credit.

  • ·Policy versioning with maker-checker controls
  • ·Salary and inflow verification signals
  • ·Audit trail tied to NBE-aligned governance
Governance

Audit-ready by design. Your credit committee stays the decider.

Every score, decision and policy change is logged with maker-checker controls, version snapshots and reason codes. DurujScore provides decision support; the lender remains accountable for the credit policy.

See governance details →Policy engine

See DurujScore on your lending workflow.

Request a demo for banks, MFIs, SACCOs, digital lenders, and merchant credit providers evaluating enterprise credit intelligence.

Request demoExplore products