Five roles, zero ambiguity
Governance formalizes accountability that already exists in business roles — rather than creating new bureaucracy. One person may wear multiple hats; one domain may hold many data products.
Data Trustee
Business · Accountable
The designated accountable authority for a specific data set, typically residing within the business. Partners with the Domain Owner on strategy and priority — not hands-on, but may step into access control, policy, and escalations.
Key responsibilities
- Endorses and is accountable for a data set within a domain
- Prioritizes data initiatives, investments, and enhancements
- Drives policy creation and sets data-quality expectations
- Resolves escalated data issues and conflicts
Domain Owner
AI Steward · Strategic
The accountable authority for an entire data domain, providing strategic oversight of all its data assets. Captures the “context” behind data to enable trusted decisions by Agentic AI and resolves cross-functional conflicts.
Key responsibilities
- Owns the domain’s data strategy and policies
- Ensures compliance, quality, and lifecycle management
- Assesses and mitigates domain risk
- Participates in the Council; advocates for the domain
Governance actions
- Collaborates with business process owners to define data access, distribution, usage, and retention for all domain data
Data Product Owner
Operational · Delivery
Provides value delivery and lifecycle management for specific data products — curated datasets, data APIs, analytics outputs. Maintains the technical connective tissue (metadata / lineage) in the AI platform.
Key responsibilities
- Defines vision, roadmap, and features for data products
- Defines and maintains metadata: definitions, PII, lineage, rules
- Ensures products meet quality and compliance requirements
- Manages access requests and security controls
Governance actions
- Accountable for data quality scores
- Ensures all subscribe-able assets have active assigned approvers
- Ensures architecture & modeling standards are followed
- Accountable for compliance: Operational, GRC, SoX, PII, GDPR, GWC
Data Steward
Upstream Business · Tactical
Drives the overall requirements for policy, regulatory compliance, and data-quality for specific data assets. Strongly aligned with the business; the primary SME for governance questions within their assets.
Key responsibilities
- Validates metadata and drives consensus on definitions
- Provides input on compliance and quality thresholds
- Leads intervention efforts when quality slips
- Maintains catalog attributes and cross-domain consistency
Governance actions
- Creates data product documentation (What, How, Business Rules) & education materials
- Identifies impacts from pipeline / lineage changes
- Reviews access requests and steers toward correct usage
Technical Custodian
Lead Data Engineer · Architecture
Designs and governs the data architecture for a domain — ensuring data is structured, stored, integrated, and used in line with enterprise standards, working closely with data architects.
Key responsibilities
- Accountable for metadata enrichment and access requests
- Owns documentation: lineage, data flows, architecture decisions
- Oversees lineage, integration, and data movement
- Evaluates domain data technologies and platforms
Governance actions
- Designs and governs data architecture & solution implementation
- Documents lineage and integration
- Implements validation rules and data-quality thresholds
RACI
DerivedAccountability across activities & roles
| Activity | Trustee | Steward | Domain | Product Owner | Custodian | Approver |
|---|---|---|---|---|---|---|
| Metadata Enrichment | I | C | I | R | AR | |
| Access | AR | R | I | RC | R | R |
| Compliance | I | C | I | A | R | |
| Quality | I | C | I | A | R | |
| Lineage | I | I | I | C | AR | |
| Modeling | I | I | I | A | R |
