Skip to main content

When something is protected

Sometimes the decision banner will say your request was redacted or blocked. That's ThreatLens protecting your organization's data — not a mistake. Here's what it means and what to do.

Redacted​

Sensitive values (like an ID number or a card number) were removed before your request was sent. You still get a useful answer — just without the sensitive details leaving.

What to do: nothing. Carry on. If you genuinely need the redacted detail in the answer, check with your administrator about your organization's policy.

Blocked​

Some content can't be shared with a model — most often secrets (passwords, API keys) or content your organization has classified as too sensitive for the chosen destination. The banner names the reason.

What to do:

  • If you attached a file that was blocked, the banner tells you which one. Try a different document, or remove the sensitive content.
  • If it's a secret, don't paste credentials into chat — that's exactly what the block prevents.
  • If you believe the content should be allowed, contact your administrator. Your organization sets these rules and can adjust them.

Why this is a good thing​

Redacted, not rejected

The goal is to let you use AI for real work while keeping a few hard guardrails. Most things go straight through; sensitive details are simply removed; and only the genuinely forbidden is blocked. You don't have to guess what's safe to share — ThreatLens handles it.

It's all on the record​

Every decision — allowed, redacted, routed, or blocked — is recorded for your organization. That record is what lets your company adopt AI with confidence. You don't need to do anything with it; it's there for your security and compliance teams.