CreditLive

A score you can explain, and a scorecard you can test first.

An explainable score from scorecards you write, with shadow runs first.

Illustrative
Why it matters

What it changes for you.

A score nobody can explain is a liability: the applicant can't challenge it, your credit team can't defend it, and a change to the model can quietly wreck the book.

Credit scoring on Bridge gives you an explainable score from 0 to 1000, worked out from behavioural features on your own platform, with the customer's consent. You write the scorecards and version them yourself. Every new scorecard runs in shadow against real applications and is back-tested before it's published, so you see what it would have done before it decides anything.

Customers can see their own score explained, see who has read their file, and dispute a score or a fact. It is not a credit bureau, and it doesn't share scores between businesses.

How it works

Step by step.

  1. Get consent

    The customer grants your business consent to score them; they can revoke it, and it expires.

  2. Extract features

    Behavioural features from their activity on your platform: income credits, savings, repayments, account age.

  3. Write a scorecard

    Your credit team writes the scorecard and its bands, versioned.

  4. Shadow-run it

    The new scorecard scores real applications silently alongside the live one, and is back-tested.

  5. Publish and explain

    Once you're satisfied, publish it. Every score comes with its reasons.

What you can switch on

Built in, not bolted on.

Scorecards you own

Your rules, your bands, your versions, not a black box you rent.

Explained to the customer

Customers see their score and what moved it, in plain words.

Shadow runs first

See how a scorecard would have decided before it decides anything.

Consent on every file

No consent, no score. Consent can be revoked and it expires.

Disputes

Customers dispute a score or a fact; your team investigates with the evidence.

Who read my file

Every read of a customer's file is recorded, and they can see it.

  • An explainable 0–1000 score
  • Scorecards you write and version yourself
  • Shadow runs and back-testing before a scorecard goes live
Use cases

How businesses use it.

Illustrative businesses, not customers. Figures are made up to show the flow.

A microfinance bank in Kano · Lenders & BNPL

Loan decisions a credit committee can defend

The bank's current model declines applicants nobody can explain, and the credit committee overrides it so often it's barely used.

  1. The credit team writes a scorecard from the factors it already trusts
  2. It shadow-runs against three months of applications
  3. Back-testing shows where it would have approved and declined
  4. Once published, every decision carries its reasons

The committee trusts the score because it can read it, and overrides become rare.

Illustrative

A BNPL provider for traders in Onitsha · Lenders & BNPL

Limits for customers with thin files

Most applicants have no bureau history at all. The provider needs a way to set a first limit from what it can see on its own platform.

  1. Traders consent to be scored on their wallet and savings activity
  2. The scorecard weights regular deposits and repayments
  3. The score sets a starting limit
  4. The limit is re-scored after each plan is repaid

Traders without a bureau file get a fair first limit, and the provider lends on behaviour it can see.

Illustrative

A digital lender in Lagos · Lenders & BNPL

A customer who disputes their score

A customer is declined and insists a repayment was recorded wrongly. Without evidence, the support team can only apologise.

  1. The customer opens their score and sees the reasons
  2. They dispute the repayment fact
  3. The lender's team investigates against the ledger
  4. The corrected fact re-scores the customer

Disputes are resolved on evidence, and a wrong fact stops costing the customer.

Illustrative
What it won't do

Honest limits.

  • It is not a credit bureau.
For developers

Read a subject's score.

Endpoints in the Credit scoring API reference, with every parameter and response.

State changes arrive as signed webhooks. How webhooks work

GET /credit-scoring/score/{subjectReference}

curl -X GET "https://api.bridgemoni.com/credit-scoring/score/b3f1c2d4-5e6a-4b7c-8d9e-0f1a2b3c4d5e" \
  -H "Authorization: Bearer $BRIDGEMONI_TOKEN" \
  -H "X-Tenant: $BRIDGEMONI_TENANT" \
  -H "X-Environment: $BRIDGEMONI_ENV"

Questions about credit scoring

Is this a credit bureau?

No. It scores customers on your own platform, with their consent. Scores aren't shared between businesses.

Can we use our own model?

You write the scorecards yourself, versioned, and shadow-run them before they decide anything.

What does the customer see?

Their score, the reasons behind it, who has read their file, and a way to dispute a score or a fact.

Bridge for Business

Tell us what you're building.

We'll come back with what's live for you today, and what's next.