Everything you need—in one package.

watsonx.governance brings together the essential capabilities for AI governance in a single, integrated platform.

Visibility

Build a complete understanding of your AI landscape. From discovering shadow AI to mapping relationships across models, risks, policies and use cases, visibility gives you the foundation to govern with confidence.

Employee support manager
AI-powered HR & IT support system
HR agent
IT support agent
Data Privacy leakage
Guardrail

Governance graph

Map connections between AI use cases, agents, models, risks, and policies for a complete picture of your governance landscape.

Employee support manager

AI-powered HR & IT support system

HR agent

IT support agent

Data Privacy leakage

Guardrail

Visibility

Governance graph

  • Know your blast radius instantly: When a model vulnerability surfaces or a policy changes, immediately traverse the full web of affected use cases, agents, controls, and obligations — across both governed and shadow AI assets.
  • Audit-ready traceability in seconds: Full governance lineage from business use case through models, risks, controls, and regulatory frameworks (EU AI Act, SR 26-2, and more), without manual assembly.
  • Surface compliance gaps proactively: Identify unmet obligations, controls lacking evidence, and non-compliant use cases mapped to specific regulatory frameworks before auditors do.
  • A governance intelligence engine, not just a visualization: The graph database enables AI-powered recommendations, impact simulation, and concentration risk analysis as the foundation for future capabilities.
3Shadow AI assets found
  • IBM watsonx Orchestrate
    Ad auction sniper
  • Amazon Web Services
    Web meeting agent
  • Microsoft Azure
    Credit agent

Shadow AI detection

Automatically identify AI assets, including Shadow AI, before they become a risk to keep a centralized, up-to-date registry of all AI assets and use cases across your enterprise.

3Shadow AI assets found
  • IBM watsonx Orchestrate
    Ad auction sniper
  • Amazon Web Services
    Web meeting agent
  • Microsoft Azure
    Credit agent

Visibility

Shadow AI detection

  • Go beyond basic catalogs: Surface the full AI system context (foundation models, tools, MCP servers, collaborator agents) to build an accurate risk profile.
  • Connect discovery to governance: Use AI-enabled semantic matching to link discovered assets to existing use cases, risks, and controls, or automatically trigger onboarding workflows for new ones.
  • Stay ahead of drift: Continuously scan for changes to agent configurations, models, and dependencies, with activity logs that keep your governance record aligned with the real AI estate.

Control

Turn insight into action with embedded governance controls. Define policies, manage risk, and align with regulatory requirements—ensuring AI systems operate within clear guardrails across their entire lifecycle.

AI risk management

Identify, assess, and mitigate AI risks throughout the lifecycle to support responsible and trusted AI adoption.

Control

AI risk management

  • Cover every dimension of risk: Enrich use case records with automatically identified risks from IBM Risk Atlas, so risk assessments are grounded in live external signals rather than point-in-time questionnaires.
  • Risk as it applies to you: Customize the platform by importing your own set of risks, so that any assessments, triaging, and workflows are specific to the problems you face in your business.
  • Risk that flows into action: External risk signals feed directly into governance workflows, dashboards, and vendor records so identified risks trigger reviews, escalations, and remediation.

Remote access policy

Description

Actions
Impact

Policy management

Create, manage, and enforce governance policies consistently across teams, models, and AI workflows.

Remote access policy

Description

Actions
Impact

Control

Policy management

  • From regulatory obligation to enforceable policy: Regulatory content providers automatically extract obligations from laws, rules, and regulations and map them to your existing policies and controls in watsonx.governance, closing the gap between what regulators require and what your policy library reflects.
  • Collaborative authoring with a full audit trail: Draft, review, and finalize policies in a structured workflow; once approved, policies are recorded in the Policy Inventory and linked to the regulatory mandates and obligations they fulfill.
  • Always aligned with current regulation: As regulatory content partners deliver updates (daily or near-real-time), changed obligations are surfaced against existing policies, triggering review workflows before non-compliance can develop.

Regulatory compliance

Align AI initiatives with leading regulations and governance frameworks to simplify compliance and reduce operational complexity.

Control

Regulatory compliance

  • Coverage across 750+ jurisdictions, 80 languages: Regulatory content partners monitor 10,000+ issuing bodies and 2,000+ regulatory sources globally, profiling content to your specific jurisdictions, topics, and firm type before it ever reaches your governance platform.
  • Horizon scanning built in: Get forward-looking intelligence on proposed and enacted rules before they take effect, giving compliance teams lead time to assess impact, update controls, and prepare evidence — rather than reacting after the fact.
  • Obligations extracted and structured on arrival: Content arrives pre-classified and obligation-extracted; mandates, sub-mandates, and requirements land directly in watsonx.governance as structured objects, not raw regulatory text that teams must interpret manually.
  • AI governance frameworks included: Purpose-built connectors for EU AI Act, ISO/IEC 42001, and NIST AI RMF ensure AI-specific regulatory obligations are treated as first-class compliance requirements alongside traditional financial services, healthcare, and data protection regulations.
Credo AI
Unified Compliance Framework

Third-party frameworks

Third-party frameworks | Access a broad ecosystem of partner-provided governance frameworks directly within the platform to extend and scale compliance coverage.

Credo AI
Unified Compliance Framework

Control

Third-party frameworks

  • An ecosystem of specialized content, not a one-size-fits-all library: Partner-provided frameworks span financial risk intelligence (RapidRatings), cyber risk monitoring (RiskRecon, SecurityScorecard), AI governance (Asenion, Credo AI), and regulatory change management (Cube, Corlytics, Wolters Kluwer, Ascent), so organizations can assemble exactly the coverage their program requires.
  • Pre-built and API-based connectors: IBM-pre-built integrations for select providers mean faster time-to-value; API-based connectors for others give flexibility to configure against your firm's specific regulatory profile, jurisdiction scope, and content preferences.
  • Scale compliance without scaling headcount: Curated external content replaces manual regulatory tracking and vendor questionnaire cycles, so compliance teams spend their time on analysis and action rather than data collection.

Embedded AI governance

Embed AI governance seamlessly into your daily activities without the overhead.

Control

Embedded AI governance

  • Seamless governance in AI development: Embed governance directly into the AI agent development workflow in IBM Bob, reducing perceived friction by making governance an automatic part of development activities.
  • Automated risk detection and assessment: As developers build or modify AI agents, watsonx.governance automatically discovers agent metadata, identifies potential risk categories, and evaluates alignment with organizational AI governance requirements, providing immediate feedback within IBM Bob.
  • Improved visibility without added developer burden: Governance teams gain structured asset metadata and risk findings, while developers receive early awareness of governance issues, increasing oversight and compliance without requiring separate governance workflows.

Accountability

Drive ROI from AI governance by embedding accountability into every decision. Automated workflows and full traceability reduce risk, accelerate approvals, and lower the cost of compliance—enabling faster, more confident AI adoption at scale.

Continuous monitoring

Continuously ingest performance, operational, and cost metrics to ensure AI systems are behaving within policy.

Accountability

Continuous monitoring

  • From policy to proof: Move beyond defining governance expectations to continuously collecting evidence that those expectations are actually being met in production.
  • Automated evidence collection: Automatically retrieve agent evaluation metrics (hallucination, helpfulness, toxicity) on a scheduled basis and store them directly as governance evidence, eliminating manual data collection.
  • Pass/breach tracking against thresholds: Evaluate collected metrics against governance thresholds and record results in real time, giving compliance leaders and auditors a single, always-current source of truth.
  • Catch issues earlier: Enable on-demand evaluation during development and continuous monitoring in production so teams can identify when agent performance drifts outside policy before it becomes a business or compliance risk.

Business value alignment

Connect AI initiatives to strategic objectives, establish measurable business cases, and continuously evaluate whether AI investments are delivering their intended value.

Accountability

Business value alignment

  • Close the ROI gap: Only 25% of AI initiatives deliver expected ROI; Business Value Alignment brings business accountability into governance so organizations can track whether AI is actually working.
  • Build business cases from day one: Define expected benefits, costs, ROI estimates, and payback calculations as part of the AI onboarding workflow, creating a measurable baseline before resources are committed.
  • Tie AI to strategy: Formally connect each initiative to enterprise priorities (Revenue Growth, Cost Efficiency, Customer Experience, Operational Excellence, and more) for portfolio-wide traceability.
  • Standardized KPIs across every initiative: Measure and compare AI performance using a common business KPI taxonomy, replacing disconnected team-specific metrics with consistent enterprise reporting.
  • Continuously monitor realized outcomes: Dashboards bring together strategic objectives, KPIs, and operational metrics in one view so leaders can compare expected vs. realized value and make informed scaling decisions.
Start
Rejected
Approved

Approval workflows

Route governance reviews through configurable stages with role-based sign-off, so use cases, models, and agents advance only with documented approval.

Start
Rejected
Approved

Accountability

Approval workflows

  • From ad hoc review to governed process: Replace email-and-spreadsheet approvals with configurable, multi-stage workflows for governed objects. Each stage creates a task, and reviewers must complete the required action before the object advances to the next stage.
  • Route work to the right people: Assign workflow tasks to users or groups, such as risk, compliance, legal, or model-validation teams, rather than relying on one named individual. Configure stage due dates, email notifications, and reminders to help keep reviews moving.
  • Turn approvals into auditable evidence: Workflow activity, including stage status and actions, is retained with the governed object’s history. This creates a traceable record of the review process that can be surfaced in task views and reports.
Use case created
Risk assesed
Model updated
Policy updated

Audit trails

Capture a system-generated record of governance decisions and changes — who did what, when, and against which version — across every governed object.

Use case created
Risk assesed
Model updated
Policy updated

Accountability

Audit trails

  • Trace decisions across connected governance records: Use object relationships to connect and navigate among governance objects, such as use cases, models, risks, controls, and policies, while retaining history at the relevant object and relationship level.
  • Capture change history as work happens: The platform records object, relationship, and workflow changes in the Activity tab. Authorized users can review the person who made a change, the time period, and—for tracked field changes—the prior and new values.
  • Find and export evidence on demand: Filter audit history by fields, relationships, workflow activity, users, and date/time ranges, then export Activity-tab data to do audit reporting whenever you need.