Signals of shadow IT
Signals are observable hints that shadow IT might exist - they’re not proof. For example, an expense line for Dropbox could mean employees are storing company files externally, but it might just as well be an external partner billing us for a shared project space. Likewise, OAuth consent to a new app could indicate a risky integration - or simply a developer testing something legitimate. Treat signals as starting points for a lightweight investigation that aims to understand the use case, data sensitivity, and risk, then either onboard the tool properly or route users to a sanctioned alternative.
The categories below highlight the most common signals that shadow IT leaves behind across different domains.
Finance
- Unapproved spend and duplicate SaaS subscriptions
- Surprise renewals slipping through unnoticed
- Invoices, vendor domains, and credit card descriptors reveal shadow subscriptions
Identity
- Accounts created outside SSO/MFA
- Password reset emails to personal inboxes
- Mismatched login providers leading to orphaned access
Data
- Sensitive files stored in unsanctioned tools
- Public links exposing information externally
- Unusual download spikes from unknown apps
Behavior
- Many users installing the same unsanctioned extension
- Survey results indicating widespread off-channel adoption
- Ticket keywords pointing to shadow usage
Technical
- Atypical TLS SNI or new SaaS endpoints in traffic
- High request volumes to unfamiliar SaaS services
- Rogue browser extensions or callbacks to unknown servers
From signal to action
The steps below outline how to move from detecting a potential signal to deciding on the right response, ensuring each case is handled with consistency and context.
-
1
Detect
Surface a signal - expense, IdP, proxy/DNS egress, API calls(example signals).
-
2
Validate
Is it real? Confirm domain/app, map to an owner/team if known.
-
3
Assess
What data? Who has access? Any integrations? Note severity & impact.
-
4
Decide
Enable with guardrails, migrate, contain, or block
-
5
Act
Notify stakeholders, implement controls, start migration if needed.
-
6
Follow-up
Capture a decision record, review after 30-60 days, update inventory.
Signal examples
Sort and filter to find the most relevant signals. Click column headers to sort. Use the chips and search to filter.