Our transaction fraud data model is designed to be a unified structure that handles both monetary (financial transfers) and non-monetary (administrative events like logins or profile changes) transaction data. This flexible model serves as the foundation to monitor all transactional and behavioural events, ensuring a consistent data structure, and hence detection capability, regardless of the event type.

Key components

At its core, every transaction includes standard identifiers for the event and the associated customer, along with a timestamp. The model supports extensive customisation through custom fields, allowing you to append key-value pairs (strings or decimals) to capture specific business context alongside the standard data.

Monetary transaction data 

For transactions involving funds, the model expands to include specialised types such as Bank Payments, Card Payments, Crypto and Remittance.

Non-monetary transaction data

Non-monetary transaction data models administrative or behavioral events that do not involve a direct transfer of funds but are relevant to enable risk detection. They utilise the same core structure as financial transactions.
Example: KYC change events signal updates to a customer's verified profile such as a change of address, identity document, or other KYC attributes - all being potential warning signs related to a customer's account.