watsonx.governance
IBM's enterprise toolkit for governing AI models end to end.
About
watsonx.governance is the governance pillar of IBM's watsonx platform. It inventories machine learning and generative AI models, generates audit-ready AI FactSheets that IBM calls nutritional labels for models, monitors them for bias, drift, and quality, and maps them to regulations including the EU AI Act, NIST AI RMF, and ISO 42001.
The toolkit combines Watson OpenScale monitoring, AI Factsheets metadata collection, and IBM OpenPages model risk governance workflows. It governs models built on watsonx.ai as well as third-party platforms like Amazon Bedrock, Microsoft Azure, and OpenAI, with cloud or on-premises deployment.
Recent additions include a Model Risk Evaluation Engine that scores foundation models against IBM's AI Risk Atlas, guardrails for generative AI applications, and a governed agentic catalog for registering and monitoring AI agents in production.
What it can do
- AI FactSheets model documentation
- Risk assessment and compliance tracking
- Bias, drift, and quality monitoring
- Foundation model risk scoring
- Governed agentic catalog for AI agents
Setup
Provision watsonx.governance on IBM Cloud or contact IBM for on-premises deployment.
Prerequisites
IBM Cloud account or enterprise IBM agreement; models to inventory.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 117; Trending score 82/100 from researcher assessment.