Fraud prevention

SOAMTAR includes a fraud scoring engine that runs before a payment request is sent to Orange Money. This page explains what you, as a merchant, can see and configure. Internal rules, thresholds and signals are not published, to keep them effective.

Scoring pipeline

POST /payments
   |
   v
[ Business checks ]        merchant limits, KYC, currency
   |
   v
[ Risk scoring ]           confidential internal rules
   |
   |-- allow   : request sent to Orange Money
   |-- review  : transaction created and flagged, awaiting review
   '-- block   : transaction refused (422)

The score and the triggered signals are stored for internal audit. You see the outcome (allow / review / block) and the decision applied, not the rules behind it.

Signal categories

For illustration only, without disclosing the exact rules, scoring covers:

• Velocity (how often a payer tries to pay)
• Blocklists and sanctions lists
• IP geolocation
• Device fingerprinting
• Amount patterns
• The payer's history with this merchant

Merchant settings

Dashboard → Security → Fraud prevention

You can:

• Add a merchant blocklist (phone numbers, emails)
• Add a merchant allowlist (skip scoring for a trusted payer)
• Set amount thresholds above which a manual review is required on your side
• Receive real-time alerts by email or webhook

Platform rules always apply on top and cannot be disabled by the merchant.

Handling a flagged transaction

When a transaction becomes is_flagged = true, you receive a dedicated event (if you subscribed to it). Recommended workflow:

1. Put the order in "review" status in your system
2. Notify your ops / risk team
3. Decide: accept or refuse from the dashboard
4. The decision is applied to the transaction automatically

Disputes

A high dispute rate leads to a manual review, then an automatic suspension if the tolerance threshold is exceeded. The exact thresholds depend on your business profile and are shared during onboarding.

To stay below the threshold:

• Automatic SMS and email receipts on every successful payment
• A clear, easy-to-find refund policy (link in the checkout)
• Time-stamped delivery logs (evidence in case of a dispute)
• Always send payer.email when you have it

Monitoring

Dashboard → Analytics → Fraud prevention

• Rate of flagged / blocked / disputed transactions
• Heatmap of risky hours (Freetown time)
• List of repeatedly flagged payers

Reporting an incident

If you suspect active fraud:

• Block the payer immediately from the dashboard
• Email support with the references involved
• SOAMTAR handles reports within the times announced by support