Protect.

Summary

Protect is a fraud detection module within the IDnow Trust Platform. It aggregates multiple signals collected during onboarding and produces detections of fraudulent behaviour, along with supporting evidence to inform decisioning.

Service Description

Protect orchestrates signals gathered by other Trust Platform components — such as document, biometric, device, email, and phone modules — and evaluates them against a set of fraud detection rules maintained by IDnow.

The service does not replace individual verification checks. It correlates their outputs to detect fraud patterns that are not visible when checks are viewed in isolation.

Protect is designed to increase detection accuracy and adapt to emerging fraud tactics without requiring Customer-side re-integration.

Key Capabilities

  • Signal aggregation

Protect collects and centralizes signals already gathered through the Trust Platform, correlating information across multiple sessions. It only evaluates rules for which the relevant signals have been collected by the Customer’s configuration, which makes it possible to identify repeated patterns and build detection rules based on them.

  • Email and Phone Number Signals

Protect uses the Email & Phone Signals component, which assesses the legitimacy and risk level of an applicant’s email address and/or phone number at the point of verification. It evaluates email quality (format validity, domain age, disposable or alias detection, spam trap flags), phone quality (number validity and type, disposable or Voice over Internet Protocol (VoIP) detection, porting history), cross-signal consistency, and network and session context. The resulting composite risk score is compared against a configurable threshold to determine the status.

  • Device Signals

Protect uses the Email & Phone Signals component, which assesses the legitimacy and risk level of an applicant’s email address and/or phone number at the point of verification. It evaluates email quality (format validity, domain age, disposable or alias detection, spam trap flags), phone quality (number validity and type, disposable or Voice over Internet Protocol (VoIP) detection, porting history), cross-signal consistency, and network and session context. The resulting composite risk score is compared against a configurable threshold to determine the status.

  • Rule-Based

IDnow’s expert team designs, maintains, and updates fraud detection rules that model known, complex fraudulent behaviours. Each rule produces a detection status when it identifies potentially fraudulent behaviour. Customers cannot create new rules, but they can configure how each rule’s detections are handled, subject to the options made available by IDnow. For example, a customer can choose to flag a session for internal review without changing its status, or to block a session when fraudulent behaviour is detected.

  • Identity Graph

Protect displays a graph illustrating relationships between onboarding sessions that share identical signal values, such as a device, a face, or a document number. This helps clients uncover connections that would otherwise remain hidden, supporting a clearer understanding of fraud patterns. These insights can also be used to feed the watchlist, which customers can then use to detect and act on similar behaviour in future sessions.

  • Watchlist

Customers may flag or block future onboarding attempts based on signal values already known to be associated with fraud. It is the Customer’s responsibility to manually add these entries to the watchlist, based on their own knowledge of fraudulent behaviour, through the interface. Some watchlist entries may be shared between Customers, to strengthen fraud detection across the IDnow customer base.

Configuration

Protect can be enabled or disabled per customer. Customers can configure and activate or deactivate the following elements:

  • Fraud detection rules
  • Watchlist entries
  • Detection thresholds

Data privacy

Data collected

Personal data processed by Protect may be shared across Customer sessions to detect fraud patterns, including cases involving multiple onboarding attempts.

The categories of personal data processed by Protect include:

  • Identity data (e.g. name, first name, date and place of birth)
  • Device data
  • Where activated: phone number, email address, and biometric data

Processing Personal Data

Protect processes signals already collected through other Trust Platform modules. Processing activities include correlation of signals, rule evaluation, score computation, and identification of related onboarding sessions.

Signals and analysis results are retained according to the retention period configured for the corresponding Trust Platform session, unless a different period is agreed contractually or required by applicable law.


Questions?