NobleCloak
Use cases

PII tables stay invisible to analysts

Author

Audrey

NobleCloak's AI Correspondent · AI-drafted, fact-checked against our sourced evidence before publishing.

Date Published

Analytics teams want their assistants close to the data warehouse. An analyst asks a question in plain language, the assistant writes the query, runs it, and returns the result. It can save hours a week, and it lowers the barrier for people who know the business better than they know SQL.

Every warehouse also holds a schema or two that analysts were never meant to see. The customer tables with names, addresses, and government identifiers are the usual example. In a warehouse with good access controls, analysts do not have them. In practice, an assistant often runs with a service credential that is broader than any single analyst's.

The problem in the data team's terms

The head of data does not want to rebuild warehouse permissions for AI. They want the rule the company already has, that analysts work from the analytics schema and the PII schema belongs to a small set of people, to hold when an assistant stands between the analyst and the warehouse.

If the assistant can query the PII schema through that shared credential, the first sign of trouble is an analyst's question that happened to be answered with data they should not have seen. The team wants the PII schema to be something the analyst's assistant cannot reach at all.

The policy

Policy is written per tool: the action, who it applies to, and one of three decisions (allowed, needs approval, denied). Role-scoped policy decides who gets a tool at all.

Action

Who

Decision

Query the analytics schema

Analysts' assistants

Allowed

Query the customer PII schema

Data Steward role

Allowed

Query the customer PII schema

Analysts and every other role

Denied, and not listed

The tool bound to the PII schema is entitled to the Data Steward role. For analysts, the assistant does not have the tool, so it cannot list the schema or query it, however the question is worded. The analytics schema remains fully available, which is where the day-to-day work lives.

Who handles it instead

When an analyst needs something that depends on PII, they ask a Data Steward, who can run the query with their own assistant or by hand and share what is appropriate. That is the same path as before AI.

If the team wants a path for occasional access, the PII tool for a role can be set to needs approval. The request then goes to a named approver or role, who answers in Slack or the web inbox. Many teams will leave it as a role entitlement because the group of people who need the tool is small and known.

What the Record of Decisions shows

The Record of Decisions has an entry per decision: who asked, which agent, which tool, the decision, who approved, and when. For the PII schema it gives the data team a short list worth reading.

  • Which Data Steward sessions used the PII tool, when, and through which agent.
  • Whether the pattern of use matches the work, such as a spike outside any known project.
  • Which analyst sessions ran against the analytics tool, so the team can see the usual traffic side by side.

If a privacy question arrives later, such as who has been querying the customer tables through AI, the answer starts from a list of known roles and sessions. It does not start from an investigation of every credential the assistants could have used.

Use it as the working record of policy decisions. The warehouse's own query history continues to hold the statements themselves.

What this does not do. It does not filter columns or rows inside a schema, and it does not change what any person can do directly in the warehouse. It controls which roles get the tool that reaches the PII schema through an assistant.

Related scenarios

Get started and scope your warehouse tools by role.