How to Configure Copilot Studio Bigquery Data Masking

Published
Length
3 min read

Luis Felipe Dias

All posts

To set up copilot studio bigquery data masking, attach column policy tags to your cloud tables. Then assign a masked reader role to the service account that your bot uses. Many teams start by linking their chat bot to the data warehouse with a shared key. When that key has broad read rights, the bot pulls clear emails and card numbers. Every user who talks to the bot can then read those raw values in chat. How copilot studio bigquery data masking works in native setups Native masking relies o

Your coding agents run with your access.

Claude Code reaches prod with a developer's credentials, and approval prompts get clicked through. hoop.dev checks every query, masks sensitive data, and holds risky changes for a human.

To set up copilot studio bigquery data masking, attach column policy tags to your cloud tables. Then assign a masked reader role to the service account that your bot uses. Many teams start by linking their chat bot to the data warehouse with a shared key. When that key has broad read rights, the bot pulls clear emails and card numbers. Every user who talks to the bot can then read those raw values in chat.

How copilot studio bigquery data masking works in native setups

Native masking relies on taxonomy tags and cloud access roles working together. First, data owners define a tag hierarchy inside the cloud catalog. They attach specific tags to sensitive columns like phone numbers and national IDs. Next, admins link a masking rule to each tag in the taxonomy. A common rule replaces the column value with a default string or a hash.

After the tags are in place, you set up the identity for the bot. The bot uses a single service account or user token to run queries. You grant that service account the masked reader role on the policy tag. When the bot sends a query, the database checks the service account role. The database engine then rewrites matched fields in results before returning the rows. The bot receives masked values for phone numbers and card numbers.

Why native roles fall short for bot queries

Native column tags solve the basic problem of static role permissions. They let you hide specific columns from a dedicated service account. Still, this setup only checks who logs in at the database level. It does not inspect what individual queries do in flight.

Most chat bots connect to the database through one shared service account. Every user in the company shares that same database identity when chatting. If you grant the bot unmasked access for one finance workflow, everyone gets it. The database cannot tell which human user asked the bot a question. Also, native tags only work on structured columns that admins tagged ahead of time. Unstructured text columns and JSON fields often hold emails and phone numbers too. If a user asks the bot to search raw text logs, tags miss them. The query runs under the valid bot role and pulls clear sensitive strings.

Data Masking (Static) + BigQuery IAM: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Managing schema changes and bot permissions at scale

Keeping column tags updated across hundreds of tables takes constant manual work. Developers add new tables and columns to the warehouse every single week. If an engineer forgets to tag a new column, the bot reads it in clear text. Auditing which columns have tags and which bots use which roles is hard.

Security teams also struggle to track what queries the bot generates on the fly. Chat bots build SQL queries based on free-form text prompts from users. A clever prompt can trick the bot into joining tables in unexpected ways. Without query-level checks, the database runs whatever valid SQL the bot sends. Teams need a way to inspect queries and mask fields based on content. Checking the actual data values in flight stops leaks when tags are missing.

Check database and API queries on the wire

Identity tools, roles, and network rules decide who opens a session. Setup alone cannot check what each query does or mask sensitive values in a response once a session starts. When workflows touch Postgres, MySQL, SQL Server, MongoDB, ClickHouse, Kubernetes APIs, SSH, or internal HTTP APIs, hoop.dev runs beside those services as an open-source proxy, a sidecar for your Postgres or API. It inspects requests coming in and results going back on the wire.

Your login details and client tools stay as they are while hoop.dev guardrails refuse bad deletes or drops before they run. On responses, hoop.dev data masking rewrites matched sensitive values in memory. Furthermore, hoop.dev holds SOC 2 Type II, is MIT licensed, and writes every query and decision to an audit trail that generates evidence for HIPAA, GDPR, and PCI DSS. See the source on GitHub.

Frequently asked questions

Does Copilot Studio mask database columns on its own?

No, the bot platform does not mask database results by itself. It passes along whatever data the warehouse returns to the active connection. You must enforce masking in the database or in your connection path.

What happens if a bot uses a shared service account?

All chat users inherit the exact same data permissions as that shared account. You cannot apply different native table rules to different chat users easily. Everyone gets the same masked or unmasked view of the data.

Can native policy tags mask emails inside free-text columns?

Native column rules apply to the entire column based on its schema tag. Redacting just an email or phone number inside a long text note requires pattern matching. Standard role tags typically replace the whole cell value instead of partial strings.

The query runs. The emails stay hidden.