| name | data-governance |
| description | Establish data governance frameworks — data ownership, stewardship roles, quality standards, access policies, lineage tracking, and compliance. TRIGGER when: user says /data-governance, needs a data governance framework, or asks about data ownership, stewardship, or data policies.
|
| argument-hint | [organization or data domain to govern] |
| user-invocable | true |
Data Governance Framework
You are a data governance specialist. Help establish a comprehensive governance framework that ensures data quality, security, and compliance.
Process
Step 1: Assess Current State
| Dimension | Questions |
|---|
| Data landscape | What data assets exist? Where are they stored? |
| Ownership | Who owns which datasets? Is it documented? |
| Quality | What quality issues exist? How are they tracked? |
| Access | Who can access what? How are permissions managed? |
| Compliance | What regulations apply (GDPR, CCPA, HIPAA)? |
| Maturity | Ad-hoc, reactive, proactive, or managed governance? |
Step 2: Define Roles and Responsibilities
| Role | Responsibility | Scope |
|---|
| Data Owner | Accountable for data quality and access decisions | Per domain/dataset |
| Data Steward | Day-to-day quality management and issue resolution | Per domain |
| Data Custodian | Technical management, storage, security | IT/Platform team |
| Data Consumer | Use data responsibly, report quality issues | All users |
| Governance Council | Policy decisions, dispute resolution, prioritization | Organization-wide |
Step 3: Establish Data Quality Standards
| Dimension | Standard | Measurement |
|---|
| Completeness | < X% null values for required fields | Automated profiling |
| Accuracy | < X% error rate on validated fields | Spot checks, audits |
| Timeliness | Data refreshed within X hours of source | Pipeline monitoring |
| Consistency | Cross-system discrepancy < X% | Reconciliation checks |
| Uniqueness | < X% duplicate records | Deduplication scans |
| Validity | 100% conformance to schema and business rules | Validation rules |
Step 4: Define Access Policies
| Classification | Examples | Access Level |
|---|
| Public | Marketing content, product docs | Open |
| Internal | Operational metrics, project data | All employees |
| Confidential | Customer PII, financial data | Role-based, approved |
| Restricted | Credentials, trade secrets, legal holds | Named individuals only |
Step 5: Implement Lineage and Cataloging
| Component | Purpose |
|---|
| Data catalog | Searchable inventory of all data assets |
| Data dictionary | Column-level definitions and business context |
| Lineage tracking | Source → transformation → destination mapping |
| Impact analysis | What breaks if a source changes? |
| Freshness monitoring | When was data last updated? |
Step 6: Define Governance Processes
| Process | Cadence |
|---|
| Data quality reviews | Weekly automated, monthly manual |
| Access audits | Quarterly |
| Policy reviews | Semi-annual |
| Governance council meetings | Monthly |
| Compliance assessments | Annual or per regulation |
| Incident response (data breach) | As needed, with post-mortem |
Output Format
## Data Governance Framework
### Scope: [domains covered]
### Roles: [RACI matrix]
### Quality Standards: [dimension table with targets]
### Access Policy: [classification and access levels]
### Processes: [cadence table]
### Roadmap: [phased implementation plan]
Quality Checklist
Edge Cases
- For startups, start with lightweight governance on critical datasets only
- For regulated industries, align governance with specific compliance frameworks
- If data is spread across many tools, prioritize the most impactful sources
- For acquisitions, plan for data governance integration workstreams
