As organizations move more workloads—including AI agents, analytical applications, and self-service data products—onto shared cloud platforms, the question of who can see sensitive data has become both more urgent and harder to answer. Traditional permission reviews and manual audits struggle to keep pace with dynamic role hierarchies, frequent data onboarding, and the expanding circle of users and services that interact with regulated information.
On October 20, 2026, Snowflake made the Sensitive Data Entitlement and Access reports generally available in the Trust Center (Enterprise Edition or higher). These reports close a critical visibility gap: one shows who can access tables containing sensitive data; the other shows who has accessed that data during a chosen lookback window. Together with Snowflake’s automatic sensitive data classification and system-provided tags, they give security, compliance, and data governance teams a practical, queryable foundation for continuous oversight.
This post examines the new reports, the underlying tagging and classification model, their compliance and operational value, integration with existing Horizon governance features, comparisons to prior tooling, implications for security and compliance teams, and what the release signals for data trust in the agentic era.
Rising Pressure Around Sensitive Data Visibility
Several forces are converging:
- Regulatory expectations (privacy, financial, healthcare) continue to demand demonstrable control over personal and regulated data.
- AI and agentic workloads increase both the volume of data access and the number of identities (human and non-human) that interact with sensitive tables.
- Cloud data platforms make it easy to share and combine datasets, which is powerful for innovation but raises the cost of blind spots in entitlement and access.
- Audit and risk teams increasingly expect evidence that is current, complete, and reproducible rather than point-in-time spreadsheets.
Knowing where sensitive data resides is necessary but no longer sufficient. Organizations also need reliable answers to two complementary questions: Who is entitled to reach it, and who has actually reached it?
Report Capabilities
The two generally available reports address these questions directly.
Sensitive Data Entitlement Report
- Lists users, roles, and privileges that grant access to tables containing sensitive data.
- Surfaces the combination of table, user, role, and privilege.
- Answers the entitlement question: “Who can access sensitive data?”
- Results are written to a view (commonly referenced under SNOWFLAKE.DATA_SECURITY) that can be queried, joined, and retained according to organizational policy.
- Supports configurable generation cadence (Daily, Weekly, Monthly, Quarterly).
Sensitive Data Access Report
- Lists users who accessed tables containing sensitive data during a configurable lookback period.
- Draws from query and access history.
- Captures user-table-role combinations, timing, frequency, and the row-access and masking policies that were in effect at the time of the most recent relevant query.
- Answers the actual-access question: “Who has accessed sensitive data?”
Both reports build on prior classification of sensitive data. Tables and columns must be identified as sensitive (via automatic classification and tagging) for the reports to include them.
The Tagging and Classification Model
Snowflake’s sensitive data classification automatically discovers and categorizes columns using system-defined tags, including:
- SNOWFLAKE.CORE.SEMANTIC_CATEGORY — Identifies the semantic type of the data (for example, name, national identifier, or other native/custom categories).
- SNOWFLAKE.CORE.PRIVACY_CATEGORY — Identifies the privacy sensitivity level of the data.
Organizations can also map user-defined tags to these system tags so that classification results automatically drive existing tagging and policy frameworks. Classification can be configured and reviewed through the Trust Center’s Data Security experience, reducing the need for manual SQL-driven discovery.
Once columns and tables are tagged as sensitive, the Entitlement and Access reports treat that classification as the scope definition. This creates a closed loop: discover → tag → report on entitlement and access → act (mask, restrict, or investigate).
Compliance Value
For compliance and audit teams, the reports provide several concrete benefits:
- Evidence generation — Queryable views replace ad-hoc permission extracts and manual access reviews.
- Least-privilege support — Entitlement results highlight over-privileged roles or users that can reach sensitive tables.
- Actual-use visibility — Access reports distinguish theoretical permissions from observed behavior, supporting both audit evidence and risk prioritization.
- Policy effectiveness insight — The Access report’s inclusion of active masking and row-access policies at query time helps demonstrate that protective controls were in force.
- Cadence and retention control — Configurable report generation and the ability to manage historical report data support different regulatory and internal review cycles.
These capabilities help organizations move from periodic, labor-intensive campaigns toward more continuous, evidence-based governance.
Integration with Existing Horizon Features
The new reports do not stand alone. They integrate with the broader Horizon governance surface:
- Automatic sensitive data classification in Trust Center (Data Security)
- System and user-defined tags
- Dynamic data masking policies
- Row access policies
- Access history and query history
- Role-based access control and privilege management
Because classification tags drive both protection (masking) and visibility (reports), governance teams can maintain a single source of truth about what is sensitive and then layer entitlement analysis, access monitoring, and protective policies on top of that foundation.
Operational Benefits for Governance Teams
Beyond compliance checkboxes, the reports deliver day-to-day operational value:
- Faster identification of toxic combinations (broad roles + sensitive tables).
- Reduced time spent assembling access evidence for audits or investigations.
- Clearer prioritization for remediation (focus first on sensitive tables that are both widely entitled and actively accessed).
- Support for data-minimization and purpose-limitation discussions by showing actual usage patterns.
- A foundation for automation—report views can feed downstream alerting, ticketing, or policy-as-code workflows.
Comparisons to Previous Governance Tooling
Before these reports, teams typically relied on a combination of:
- Manual or scripted privilege queries against the access-control hierarchy
- Periodic access reviews driven by spreadsheets or third-party tools
- Custom joins between classification results and access history
- Point-in-time exports that became stale quickly
The GA reports improve on that baseline by:
- Scoping automatically to classified sensitive data rather than requiring manual table lists
- Producing ready-to-query views on a chosen cadence
- Combining entitlement (potential) and access (actual) perspectives in a consistent Trust Center experience
- Incorporating policy context (masking and row access) into access results
The result is less custom engineering and more standardized, repeatable visibility.
Implications for Security and Compliance Teams
Security and compliance teams gain a practical control that matches the reality of modern data platforms. Instead of treating entitlement and access as occasional projects, they can treat them as ongoing, queryable properties of the environment. This supports:
- More frequent, lower-effort access reviews
- Better alignment between data classification and access governance
- Stronger evidence packages for internal and external auditors
- Clearer input into risk acceptance and exception processes
- A shared language between data platform, security, and compliance stakeholders
In an environment where AI agents and automated workloads also access data, the ability to see both human and service-principal activity against sensitive tables becomes increasingly important.
What This Signals for Data Trust in the Agentic Era
As organizations deploy agents that reason over and act upon enterprise data, trust in the underlying data platform becomes a prerequisite for scale. Agents amplify both the value and the risk of broad data access. Features that make sensitive data location, entitlement, and actual usage visible and governable are therefore foundational infrastructure for the agentic era—not optional compliance add-ons.
The GA of Sensitive Data Entitlement and Access reports signals that Snowflake is treating continuous, classification-driven visibility into sensitive data access as a core platform capability. Combined with automatic classification, tagging, masking, and row-access policies, these reports help close the loop between discovery, protection, and oversight.
Actionable Insights for Governance Teams
- Ensure sensitive data classification is enabled and healthy for the databases that matter most; the reports are only as complete as the classification scope.
- Enable the Entitlement report on a cadence that matches your risk profile (weekly or daily for highly regulated domains).
- Run Access reports with lookback windows aligned to audit or investigation needs, and retain results according to your evidence-retention policy.
- Query the entitlement view for high-privilege roles that reach many sensitive tables and prioritize review of those combinations.
- Compare entitlement versus actual access to identify both unused privileges (candidates for reduction) and unexpected access patterns (candidates for investigation).
- Feed report outputs into existing ticketing or workflow systems so that findings drive remediation rather than remaining as static reports.
- Align tag maps so that classification results automatically drive your organization’s standard sensitivity labels and policy frameworks.
- Include non-human identities (service users, agents) in your review processes; the reports surface access regardless of identity type.
- Treat the reports as living controls: schedule periodic review of cadence, scope, and downstream processes as data and usage patterns evolve.
- Coordinate with platform and AI teams so that new agentic workloads are onboarded with clear expectations about sensitive-data access and monitoring.
Sensitive Data Entitlement and Access Reports GA gives organizations a practical, integrated way to answer two of the most important questions in modern data governance: who can reach sensitive data, and who has actually reached it. By grounding those answers in automatic classification and system tags, and by delivering results as queryable views inside the Trust Center, Snowflake has strengthened the foundation of data trust required for both classic analytics and the emerging agentic enterprise. For governance, security, and compliance teams operating in 2026, these capabilities turn visibility from a periodic project into a continuous, actionable control.