> ## Documentation Index
> Fetch the complete documentation index at: https://help.proctorly.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Verdicts & actions

> Confirm or dismiss individual flags, set a session verdict of Clear or Escalate, and clear sessions in bulk.

After reviewing a session's flags and integrity score, you record your decisions. Proctorly has two levels of decision: **per-flag** (confirm or dismiss each event) and **per-session** (a verdict that can override the whole score).

## Per-flag decisions

For each flag in a session, you choose:

<CardGroup cols={2}>
  <Card title="Confirm" icon="check">
    You consider the flag genuine. It counts at **full weight** toward the integrity score.
  </Card>

  <Card title="Dismiss" icon="x">
    You consider the flag a false positive or not worth counting. It **stops counting** toward the score.
  </Card>
</CardGroup>

A flag you haven't decided on yet counts at a **reduced weight** — so leaving everything undecided isn't the same as dismissing it. Proctorly can auto-suggest decisions based on confidence (see [Reviewing results](/teachers/reviewing-reports#decisions-confirm-or-dismiss)).

## Session verdict

A verdict applies to the **whole session** and overrides the per-flag math:

<CardGroup cols={2}>
  <Card title="Clear" icon="circle-check">
    Forces the integrity score to **100**, regardless of any flags.
  </Card>

  <Card title="Escalate" icon="triangle-alert">
    Treats **every** flag — even undecided ones — at full weight.
  </Card>
</CardGroup>

<Info>
  These are the **only two verdicts**. Use **Clear** when a session is fine despite some flags, and **Escalate** when it needs to be treated as a serious case.
</Info>

## Clearing sessions in bulk

When many sessions are clearly fine, you can clear a set of them at once rather than opening each one. This applies the **Clear** verdict across the selected sessions in a single action.

<Info>
  **Reviewer TODO.** The exact in-app control for bulk-clearing (its label, and how sessions are selected — for example by risk level) should be confirmed with the product team before documenting the precise steps.
</Info>

## A note on tone

Flags are detections, not conclusions. The integrity score and verdicts are tools for **your** judgment — Proctorly records what it observed and leaves the determination to you. Any AI-written summary is assistive only; see [Reviewing results](/teachers/reviewing-reports#ai-summaries).