26 DLP rules (Drive/Gmail/Chat)

Data protection rules scan Drive files, Gmail messages and Chat messages for content that matches a detector — card numbers, national IDs, credentials, or a custom regex — and audit, warn or block when that content is shared or sent. A rule is scoped by OU or group, by app, by trigger condition and by detector, and its alerts land in the alert center at the severity you choose. The usual sequence is audit-only to size the false positives, then warn, then block.

Documentation: About DLP ↗

Caveats

A rule that only logs is a detection, not a prevention¹
Scan-on-share triggers on the sharing event, not on content at rest²
Detectors are regex over extracted text³
DLP requires Enterprise Standard+, Frontline Standard+, an Education edition or Enterprise Essentials Plus⁴

Setup steps

open ↗

https://admin.google.com/ac/dp · captured 2026-07-15

Security › Access and data control › Data protection
Open Data protection and go to Manage rules; review the existing rules and detectors before adding more — duplicate rules with conflicting actions are the usual failure mode.
open ↗

https://admin.google.com/ac/ax?action_id=RULE_TEMPLATES · captured 2026-07-15

Security › Access and data control › Data protection › Add rule > New rule from template
Start from a template rather than a blank rule for the standard detectors (credit card, national ID, credentials) — under Manage rules choose Add rule > New rule from template and pick the template matching your classified data types.

Create data protection rules ↗
open ↗

https://admin.google.com/ac/dp/createrule · captured 2026-07-15

Security › Access and data control › Data protection › Add rule > New rule
Scope the rule: choose the OU (organizational unit) or group, the apps (Drive, Gmail, Chat), the trigger conditions and the content detectors.

Scope

all users

Apps

Drive + Gmail + Chat

Conditions

content matches detector

Trigger

file shared externally / message sent externally

Create data protection rules ↗ Create DLP for Drive rules and custom content detectors ↗
Security › Access and data control › Data protection › Add rule > New rule
Set the action and the alerting. Run in audit-only mode first to size the false-positive rate, then escalate to block.

Action = Audit only (week 1) → Warn → Block (Drive: Block external sharing; Gmail: Block message, outgoing messages only); Alerting severity = High, send to the alert center

Ongoing maintenance

automatable: AI agent Monthly: review DLP alert volume and false positives; tune detectors and thresholds.
requires a human Per new data type the org handles: extend detectors and re-run the canary test.

How to verify

Create a canary document containing a synthetic detector hit (test credit-card number), share it externally from a test account, and confirm the rule blocks/warns and raises the alert.

v0.1.2 Prevent edition Ent Std+ policy #15 · #33 ↗