CHECKLIST, CONDITION 10 OF 10, read 13 Nov 2026
Access policy allows AI use: 7 of 17 tools check it by rule, 6 leave it to a person
- The test
- An access policy exists and allows this agent to use the asset for this purpose, for example approved for AI.
- Why it matters
- An agent often runs on a broad service account. Without an explicit policy it can read or change things no person would be allowed to.
- How to check it
- Check the asset policy, the agent role and any approved-for-AI tag. Confirm the agent account can reach only the assets approved for its use.
- What failing looks like
- The agent runs under an admin account and reads HR salary tables nobody approved for AI.
What each use case needs
| Use case | Need | Why |
|---|---|---|
| Metric agent | required | The agent must only read what the asking user is allowed to see. |
| Coding agent | required | A coding agent should have only the repository and warehouse permissions its task needs. |
| Write-back agent | required | Write access is the highest-risk permission and must be granted on purpose, per target. |
7 tools check access policy allows AI use by rule
- Euno, rank 2, 62.5.
- Atlan, rank 5, 59.0.
- Databricks Unity Catalog, rank 7, 58.0.
- dbt, rank 8, 52.5.
- Snowflake Horizon, rank 13, 45.5.
- Informatica, rank 16, 36.0.
- Collibra, rank 17, 36.0.
6 tools leave access policy allows AI use to a person
- OvalEdge, rank 1, 64.5.
- DataHub, rank 3, 61.0.
- OpenMetadata, rank 4, 61.0.
- Secoda, rank 6, 58.5.
- Ataccama, rank 10, 49.5.
- Actian, rank 15, 39.0.
4 tools show none found on 13 Nov 2026
- Alation, rank 9, 50.5.
- Qlik Talend Cloud, rank 11, 46.5.
- Monte Carlo, rank 12, 46.0.
- Precisely, rank 14, 43.0.
What the sources say
- Atlan, AI readiness assessment for data estates: "Agents running on broad-read service accounts are the most common way a well-governed estate leaks"
- Gartner, AI-ready data: what it is and how to get there (Rita Sallam; archived copy): "Data stewardship: Apply policies throughout the data life cycle, including model access and development."
- GO FAIR Foundation, The FAIR Guiding Principles: "(meta)data are released with a clear and accessible data usage license"
Other conditions
Sources
- Atlan: AI readiness assessment for data estates, read 13 Nov 2026
- Gartner: AI-ready data: what it is and how to get there (Rita Sallam; archived copy), read 13 Nov 2026
- GO FAIR Foundation: The FAIR Guiding Principles, read 13 Nov 2026
Last reviewed 13 Nov 2026.