IBM watsonx.governance

AI Model Inventory

Centralized registry of all AI/ML models in use across the organization, with ownership, lifecycle stage, and risk classification.

Model Monitoring

Production monitoring for performance, drift, data quality, and fairness regressions.

Bias & Fairness Testing

Automated statistical testing for disparate impact across protected attributes, with audit-ready reports.

Explainability

SHAP, LIME, counterfactual, and feature-importance explanations for model decisions.

Drift Detection

Automated detection of distribution shift, feature drift, prediction drift, and performance degradation in deployed ML/AI models.

Risk Assessment Workflow

Guided workflows for completing AI impact assessments, risk scoring, and approval routing.

Policy Management

Authoring, versioning, and distribution of AI usage policies mapped to regulations.

Audit Logging

Tamper-evident logging of governance events (approvals, model changes, policy decisions) required by EU AI Act Article 12 and similar regulations.

Regulatory Intelligence

Ongoing tracking of AI laws, standards, and enforcement, mapped to your model inventory.

Hallucination Detection

Automated scoring of LLM outputs for factual accuracy and groundedness, typically via retrieval-backed verification or reference comparison.

LLM Guardrails

Runtime input/output filtering for LLMs — PII redaction, toxicity blocking, prompt injection defense, policy enforcement.

LLM Evaluation

Systematic testing of LLM outputs for correctness, relevance, safety, and consistency using automated scorers, rubrics, or human review.