# Risk AI Agent

The Risk AI Agent module adds [Risk AI Agent](https://developer.incode.com/docs/risk-ai-agent), Incode's intelligent fraud decisioning system, to a verification session. It evaluates signals from across the session and produces a more accurate pass/fail decision than set thresholds or custom rules.

## Integrations

:white_check_mark: iOS | :white_check_mark: Android | :white_check_mark: Web SDK

## How It Works

Risk AI Agent runs as a Process node in a Flow or Workflow. It is designed to be placed at the end, after upstream modules, such as [ID Validation](https://developer.incode.com/docs/id-validation-module), have produced their outputs. Because Risk AI Agent evaluates all available session signals holistically, it performs best when those signals are already present. Adding it mid-flow, before enough signal has been generated, can produce unreliable results.

When Risk AI Agent runs, it replaces the legacy TotalScore model as the final session decision mechanism. Rather than returning a numeric score, it outputs a pass/fail result directly. [Single Session view](https://developer.incode.com/docs/single-session-view#risk) shows the Risk AI Agent pass/fail decision and the top 3 contributing signals for that session.

## Use Risk AI Agent

- [Risk AI Agent (Dashboard)](https://developer.incode.com/docs/risk-ai-agent-dashboard)
- [Risk AI Agent (iOS)](https://developer.incode.com/docs/risk-ai-agent-ios)
- [Risk AI Agent (Android)](https://developer.incode.com/docs/risk-ai-agent-android)
- [Risk AI Agent (Web SDK)](https://developer.incode.com/docs/risk-ai-agent-web-sdk)

<br />