A score should open a conversation, not close a door.

CreditPrism combines transaction, telco and behavioural signals to help lenders examine thin-file borrowers with more context—and communicate how a result was formed.

Read the thesis

The thesis

Absence of a traditional file is not absence of financial behaviour.

A salary pattern, airtime history or steady bill rhythm can add useful context where conventional bureau data is limited.

CreditPrism is designed as an evidence layer: it organises permissioned signals, highlights their influence and gives an underwriter a reasoned view rather than a mysterious number.

The aim is not to remove human judgement. It is to make that judgement better informed, more consistent and easier to review.

Interactive score model

Move the signals.
See the reasoning.

This educational sandbox uses fictional inputs and an illustrative formula. It is not a lending decision or a real score.

Illustrative score 671 Developing

Primary influence / 01

Telco continuity is the strongest signal in this sample profile.

Signal architecture

Three lenses.
One contextual view.

TX / 01

Transaction patterns

Read cash-flow regularity, balance behaviour and recurring commitments from permissioned transaction data.

TC / 02

Telco continuity

Examine tenure and usage continuity as supporting context, with relevance and consent kept visible.

BH / 03

Behavioural consistency

Bring steady financial actions into view without pretending every behaviour predicts credit readiness.

Governance by design

Useful intelligence needs visible boundaries.

Permission before analysis

Make the source, purpose and consent status of each signal understandable.

Reasons alongside results

Show the factors influencing an assessment so it can be reviewed and challenged.

Human review stays present

Route uncertain or sensitive cases into a clearly defined manual workflow.

Monitor the model

Give teams space to inspect drift, outcomes and signal relevance over time.

Continue the enquiry