THE REVIEW WINDOW — BEFORE ANYTHING IS SENT
1
Sensitive content is detected in the document and highlighted in place, so the user sees exactly what was found and where.
2
Protected items can be reviewed and corrected — an incorrect detection can be removed, and a missed value can be marked manually.
3
The user explicitly continues to the AI workflow — nothing is sent until the review is confirmed.

Why review exists#

Detection is layered and tunable, but no detector should be presented as perfect. Human review is the architectural answer to the question every security team asks — what if the model misses something? Before content leaves the environment, the person who actually understands the document gets to see what was detected, fix what was wrong, and add what was missed. The correction happens at the moment it is cheapest: before egress, not after an incident.

Where review appears#

Human review exists in both product experiences where configured:

  • Salus Desktop — shows the review window for files: when a document is about to cross the boundary, detected items are visible and correctable, including manual marking of items the detector missed. Typed and pasted text is tokenized automatically without a review step.
  • Salus API — in the integrated path, review belongs to your application: Salus outputs the protected content and the detected items, and your workflow presents its own approval step where regulation or policy requires a human decision before egress.

Review is a configurable control, not a mandatory gate: most workflows run straight through with automatic tokenization, and policy decides which classes, users, or destinations warrant the extra step. That keeps the protection proportionate — friction where the risk justifies it, invisibility everywhere else.

What review is for#

Review is the architectural complement to detection, not a patch on it: the person who understands the document confirms the protection at the moment correction is cheapest — before egress, not after an incident. In integrated deployments this also matches how regulators frame it: the final human say lives in the customer's own workflow, on the customer's own UI.