Fraud Intelligence
Fraud Intelligence uses AI-powered machine learning models to automatically assign risk scores to card-not-present authorization and verification transactions, and to detect fraudulent patterns before they result in losses. The outcome of the evaluation is returned in the response using the risk.riskEvaluationCode field, which indicates whether the transaction continued to the issuer or was stopped by Fraud Intelligence. You do not need a separate API call.
Before you begin
Fraud Intelligence is enabled and managed in Commerce Center. No API changes are required to your authorization or verification integration to get started. For details on configuring rules and allow and deny lists, refer to the Commerce Center Fraud Intelligence user guide.
Availability
| Region | Status |
|---|---|
| United States (US) | Supported |
| Canada (CA) | Not yet available |
| EMEA (UK) | Not yet available |
| Australia (AU) | Not yet available |
Out of scope
Fraud Intelligence does not currently evaluate:
- Non-card payment methods
- Card-present transactions
Request fields that drive Allow and Deny lists
Allow and Deny lists are merchant-configured lists of attributes that immediately trust or block matching transactions. Fraud Intelligence matches Allow and Deny list entries against specific fields on your authorization requests. If you don’t include a field in the request, the list rule for that attribute cannot match.
| Commerce Center attribute | API field name | Format |
|---|---|---|
| Customer Email | accountHolder.email |
Valid email format. Must contain an @ and a period (.). |
| Customer ID | accountHolder.referenceId |
Must match the exact value you configure on the list. |
| IP Address | browserInfo.deviceIPAddress |
Valid IPv4 or IPv6 address. |
API behavior in responses
When Fraud Intelligence is enabled, every authorization and verification response includes a risk.riskEvaluationCode field. Read this field to see the outcome of the fraud evaluation for the transaction.
| Field name | Data type | Description |
|---|---|---|
risk.riskEvaluationCode |
string (enum) |
The result of the fraud evaluation for the payment request. Indicates the next steps in the payment processing flow.
|
Code samples
The following examples show portions of authorization responses in which Fraud Intelligence approved or denied a transaction.
Approve
The following code sample shows a portion of an authorization response in which the transaction continued to the issuer..
HTTP Method: POST
Endpoint: /payments
{
"transactionId": "v2-auth-1781800001",
"requestId": "auth-req-1781800001",
"transactionState": "AUTHORIZED",
"responseStatus": "SUCCESS",
"responseCode": "APPROVED",
"responseMessage": "Transaction approved",
"amount": 15000,
"currency": "USD",
"paymentMethodType": {
"card": {
"cardTypeName": "VISA",
"maskedAccountNumber": "401200XXXXXX0026"
}
},
"risk": {
"riskEvaluationCode": "APPROVE"
}
}Deny
The following code sample shows a portion of an authorization response where Fraud Intelligence denied the transaction, stopping the payment before it was sent to the payment network. The risk.riskEvaluationCode field shows the result of the fraud evaluation, reflected in the responseCode and responseMessage.
HTTP Method: POST
Endpoint: /payments
{
"transactionId": "v2-auth-1781800002",
"requestId": "auth-req-1781800002",
"transactionState": "DECLINED",
"responseStatus": "DENIED",
"responseCode": "RISK_DENIED",
"responseMessage": "Transaction was declined due to fraud check",
"amount": 95000,
"currency": "USD",
"paymentMethodType": {
"card": {
"cardTypeName": "VISA",
"maskedAccountNumber": "401200XXXXXX0026"
}
},
"risk": {
"riskEvaluationCode": "DENY"
}
}Risk factor guidance
Fraud Intelligence evaluates every transaction using the data your integration sends. The more relevant data you include, especially device and identity data, the more accurately it may distinguish legitimate consumers from fraud.
Risk data by group
Send as much data from each group as your integration supports. The groups are ordered by impact, starting with the highest-impact group.
| Data group | What to send | Why it matters |
|---|---|---|
| Device signals | browserInfo (send all fields) |
Provides a strong signal for detecting card testing and bots. |
| Customer identity | accountHolder (send all fields — e.g. firstName, lastName, email, billingAddress) |
Identifies the shopper and links repeat customers across transactions. |
| Ship-to (physical goods) | shipTo (send all fields — e.g. shippingAddress) |
A mismatch between billing and shipping is a classic fraud signal. |
| Order context | merchantOrderNumber, initiatorType, accountOnFile, recurring (send all fields) |
Provides the model with purchase context, such as whether the transaction is recurring, who initiated it, and whether a saved card was used. |
| Merchant | merchant (send all fields — e.g. merchantSoftware, merchantCategoryCode) |
Attributes the transaction to your integration. |
It gets better the longer you send it. The more consistently you submit this data, the better Fraud Intelligence can learn your business’s normal patterns and make accurate, tailored decisions.
Related
Fraud Prevention
Fraud Intelligence
Online payments error codes