Verdict
Disposition, confidence, and a recommended action: block and freeze, hold for review, contact the customer, release, or no action.
Fraud investigation API
Lumin Reveal sits beside the fraud system you already run. Post an alert. The service investigates and returns a verdict, a narrative, a complete timeline, and an itemized cost.
How it works
Your fraud system posts an alert. The API accepts it at
POST /v1/investigations and returns an investigation
ID immediately.
The service gathers customer context, reviews transaction activity, and checks accounts, devices, and IPs against threat intelligence. When the case warrants it, it interviews the customer by voice, SMS, or WhatsApp, in the customer’s language.
A typed decision model returns the disposition, a confidence, and a recommended action. Read the result from the API, or receive it at a callback you provide.
The record
Every completed investigation comes back as a record an investigator, a model-risk reviewer, or an auditor can read.
Disposition, confidence, and a recommended action: block and freeze, hold for review, contact the customer, release, or no action.
A written account of the case and the findings that drove the decision.
Every step, decision, and tool call, in order. The audit trail stays with the investigation.
An itemized cost of the investigation. Exact and estimated amounts are reported separately.
Outcomes
The investigation still finishes when the customer cannot be reached, an upstream system is unavailable, or the budget for the case is spent. Inconclusive is a completed result.
The evidence supports fraud.
The evidence points to fraud and is not conclusive.
The alert is explained by legitimate activity.
The evidence on hand does not support a stronger call.
Authorization under coercion is not treated as a legitimate approval. Customer contact is used only when the case warrants it.
Contact
For fraud, operations, and technology teams evaluating an investigation API.
hello@luminreveal.com