MRM P&C - Explainability and Transparency Engine

Purpose

MRM P&C - Explainability and Transparency Engine supports model governance by computing interpretability metrics such as SHAP and LIME and using GenAI to explain them in plain language for governance reports.

Primary users

The primary user is not specified. Based on the provided use case, the agent is intended for users involved in explainability and transparency activities for insurance model governance reports.

Where it fits (process/stage/trigger)

The agent fits into a governance reporting process after model input variables, output probabilities, and explainability metrics are available and need to be translated into clear explanations.

Key capabilities / workflow

The agent computes interpretability metrics including SHAP and LIME, checks whether the metrics are available, generates plain-language explanations with GenAI, and supports completion of governance report content.

Inputs

Explicit user inputs are not specified. The provided dataset context includes model input variables, output probabilities, and explainability metrics.

Outputs / Deliverables

Explicit outputs are not specified. The provided use case states that the agent produces plain-language explanations of interpretability metrics for governance reports.

Value

The agent helps improve explainability and transparency by turning technical interpretability metrics into understandable language for insurance governance reporting.