Skip to main content

Policy Tasks and the Policy Audit Trail

Find open policy tasks, review scan feedback, and see every change made to your AI Policies, including applied change sets.

Written by Austin Carroll

The Policy audit trail page is where admins keep their AI Policies healthy. It shows how your team has rated scan findings and a full record of every policy change, including changes applied from the AI assistant. This article also explains policy tasks: the open work Warrant raises about your policies.

Where to find it

  1. In the left sidebar, click Settings.

  2. Under Compliance, click Policy audit trail.

You can also search Settings for "audit", "history" or "change set". Older links to the audit trail or the policy tasks page now open this page.

What is a policy task?

A policy task is a suggestion that one of your policies needs a look. Warrant creates tasks automatically, so you don't need to set anything up. When a task is raised, you get a notification, and open tasks about a policy are listed under Open tasks in the How it's doing section of that policy's page.

You'll see these kinds of tasks:

  • A policy is frequently marked not relevant. When reviewers repeatedly mark findings from the same policy as not relevant, Warrant suggests you review that policy. It may be too broad for your content, or not apply to your business at all. The task names the most common reason reviewers gave, when they gave one.

  • Warrant updated a regulation you use. When Warrant updates one of its built-in regulations that you have turned on, your scans keep using the version you have until someone reviews and accepts the update. Clicking the notification opens the regulation's page, where a Warrant updated this regulation banner lets you click Review changes to compare your version with Warrant's, then Accept update. Once accepted, scans use the new version.

  • A reviewer added a flag the scan missed. When a reviewer uses Add flag on a scan, a task named Reviewer flag: followed by the flag's title is raised, showing its severity, the file, and the policy the reviewer chose (or "no policy"). Use it to decide whether a policy needs to change, or whether you need a new one. The task is for your team to review: a reviewer flag never changes a policy by itself. See Working with Scan Findings.

Super admins are notified about feedback and reviewer flag tasks. Notifications about Warrant's regulation updates go to everyone who can manage AI Policies.

Work a task

  1. Open the policy from the notification, or find it on the AI policies page and click it.

  2. Check How it's doing for the votes, the reasons reviewers gave and their notes, and the open tasks.

  3. Make any changes you need. For example, edit the policy, narrow where it applies, lower its risk level, or turn it off. For a Warrant update, review and accept it from the banner at the top of the page.

Scan feedback summary

When reviewers give thumbs up or thumbs down on scan findings, the Scan feedback summary shows the totals for each policy, with Relevant and Not Relevant counts. Expand a policy to see the individual findings and votes. Votes are counted per policy, for your company only. Repeated "not relevant" votes on a policy make Warrant apply it more strictly, and can lead to its new findings being dismissed automatically.

  • Search by policy, or filter by regulator or by Mostly Relevant / Mostly Not Relevant.

  • If Warrant is applying some policies more strictly because of your feedback, a banner lists them. Click a policy name to open it.

  • Click Download Report to export the summary as a CSV file.

What the policy audit trail records

The Policy audit trail is a chronological log of policy changes, newest first. Each entry shows who made the change, what they did, the policy it applied to, any note they left, and the date. Entries include:

  • Policies added, edited, deleted, or customized

  • Policies enabled or disabled, one at a time or in bulk

  • Risk level changes, including the old and new level

  • Policies restored to their default

  • Attachments added to, deprecated on or restored to a policy

  • Warrant's regulation updates your team accepted

  • Relevancy feedback on scan findings

  • Reviewer flags added to scans, marked Reviewer flag, with the flag's title and severity under the policy the reviewer chose (or "no policy")

  • Changes applied from the AI assistant, marked Change set, with the change set's name and the fields that changed. If you can use the AI assistant, click the entry (View details) to open the change set and see each change, or undo it.

Changes made automatically by Warrant show System as the user.

Filter and export the audit trail

  1. Narrow the list with the filters above it: All users, All policies, All Actions (Enabled, Disabled, Added, Edited, Deleted, Customized, Restored, Risk Level, Feedback, Reviewer flag or Change set), and a From / To date range.

  2. Click Clear to reset the filters.

  3. Click Download CSV to export the entries you're viewing. The file includes the date, user, action, policy and note for each entry. For a reviewer flag, the note names the flag and its severity. For a change set entry, it names the change set and the fields changed.

The export follows your filters, so set them first if you only need a specific period or policy for an audit. The page shows the most recent 200 policy changes.

Who can use it

This page is available to admins, including people with a capability-based admin role such as Compliance Admin. Accepting Warrant's regulation updates and opening change sets needs the Manage AI Policies permission, which admins have unless your company gives it only to certain roles. Giving thumbs up or down on scan findings needs the Relevancy Feedback permission.

Did this answer your question?