Skip to content

Foundry Sentinel · AI-Powered Fraud and Anomaly Monitoring

Identify unusual transaction activity. Give investigators better visibility.

Foundry Sentinel helps financial institutions identify unusual transaction patterns and bring potentially suspicious activity to the attention of investigators. Combining transaction monitoring approaches with machine-learning techniques, Sentinel is designed to help teams examine patterns, understand alerts and make more informed investigation decisions.

Flagged

Why it was flagged

Investigate Dismiss

The challenge

Fraud monitoring needs to keep pace with changing transaction behaviour.

Banks operate across multiple transaction channels and face evolving fraud patterns. Monitoring teams need to identify suspicious activity, distinguish meaningful alerts from normal behaviour and investigate cases efficiently.

A monitoring environment that provides limited context can make investigation more difficult.

Foundry Sentinel is designed to improve visibility into unusual activity and support the work of fraud monitoring teams.

The solution

Move from transaction alerts to better-informed investigations.

Foundry Sentinel analyses transaction activity to identify patterns that may warrant further investigation.

It is designed to help investigators understand why an alert was raised, examine the relevant transaction context and record their assessment.

The objective is not simply to generate more alerts. It is to help teams focus their attention and improve the quality of investigation.

How it works

How Foundry Sentinel works

  1. Transaction data
  2. Pattern analysis
  3. Alert
  4. Investigation
  5. Feedback
  1. 01Monitor transaction activity

    Assess transaction data from agreed channels and sources.

  2. 02Identify unusual patterns

    Use configured rules and machine-learning techniques to identify activity that differs from expected patterns.

  3. 03Generate explainable alerts

    Present relevant indicators and transaction context to help investigators understand why activity has been flagged.

  4. 04Investigate and assess

    Enable investigators to review alerts, examine available information and record their findings.

  5. 05Improve monitoring

    Use investigation feedback and monitoring outcomes to inform ongoing review and refinement of detection approaches, where supported by the implementation.

Capabilities

What Foundry Sentinel does.

  • Transaction pattern monitoring

    Identify unusual activity across agreed transaction data and monitoring scenarios.

  • Rules and machine-learning techniques

    Combine configured monitoring rules with machine-learning approaches to identify patterns requiring attention.

  • Explainable alerts

    Provide relevant context and indicators to help investigators understand why activity has been flagged.

  • Investigation workflow

    Support the review, assessment and handling of monitoring alerts.

  • Investigator feedback

    Capture investigation outcomes to support monitoring review and refinement.

  • Operational visibility

    Provide a structured view of alerts and investigation activity.

Use cases

Monitoring scenarios

Foundry Sentinel can be evaluated against relevant monitoring scenarios, such as:

  • Unusual transaction frequency or value
  • Deviations from established transaction patterns
  • Potentially suspicious transaction sequences
  • Unusual activity across agreed customer or account dimensions
  • Transaction activity requiring additional investigation

The scenarios and detection criteria should be agreed with the institution based on its risk priorities and available data.

Designed for controlled evaluation

Fraud monitoring solutions must be assessed carefully. Data quality, detection coverage, alert quality, explainability and integration with existing processes all matter.

Foundry Sentinel can be evaluated through a controlled engagement with agreed transaction data, monitoring scenarios and assessment criteria.

FAQ

Common questions.

Does Foundry Sentinel replace our existing fraud monitoring system?

It should be evaluated in the context of your existing monitoring environment. The objective is to assess where additional detection and investigation capabilities may add value.

Does it use machine learning?

The solution is designed to combine machine-learning techniques with configured monitoring rules. The specific approach depends on the use case and implementation.

Can investigators understand why a transaction was flagged?

Explainability is an important part of the product proposition. Alerts are intended to provide relevant indicators and context for investigation.

Can it monitor different transaction channels?

The channels and data sources to be evaluated are agreed during solution scoping.

Is Foundry Sentinel generally available?

Contact our team to discuss its current availability and early-access evaluation arrangements.

Explore Foundry Sentinel

Discuss your current fraud monitoring environment and identify where additional pattern analysis and investigation visibility could help.