Skip to content
English
  • There are no suggestions because the search field is empty.

Notes: June 8th Release

This release introduces AI-powered item review and audit tracking, UI cleanup in authoring, improved visibility of AI feedback in items, extended proctoring recordings after termination, and updated Certiverse branding on the B2C login page.

Please note: In the interest of informing our customers as early as possible of any platform changes, we are sharing these release notes in advance of the software release. The final updates may vary and will be documented accordingly.

Home

Update B2C Login Page with New Certiverse Branding

The Microsoft Entra B2C login page, displayed when users click "Log in" on certiverse.com has been updated to reflect new Certiverse brand elements. All visual assets and styles, including the background graphic pattern and layout, have been redesigned according to approved Figma mockups to deliver a consistent, modern brand experience from the very first touchpoint. No functional changes to the login or authentication flow were made.

Authoring

Remove Redundant Thumbnail Videos and MQC Popups from Authoring Screens

Thumbnail video components have been removed from the JTA screen and the Item Writing → View/Edit Item Assignment screen, along with the Minimally Qualified candidate (MQC) popup that previously appeared after clicking "Got It" on the JTA screen — including its "Do not show this window again" checkbox and functionality. The same checkbox has also been removed from the MQC modal accessible via the header MQC button.

The Guide button remains as the sole, standardized entry point for instructional content across all affected screens, eliminating duplicated guidance and visual clutter.

Blueprint

AI-Powered Item Review Step

When the Enable AI Item Review toggle is turned on for a Blueprint, single correct response items (MC2–MC6) without images are now automatically reviewed by Certiverse AI as the first step in the review process upon submission. The AI evaluates four review dimensions — stem clarity, key correctness, key distinctiveness, and distractor quality — and returns an overall outcome of Clearly Pass or Clearly Fail, which drives the item through the existing review workflow just as a human reviewer would.

Items containing images, non-eligible item types, or items that have already received an AI review bypass this step and follow the standard human review flow. If the AI review service fails, the item is automatically routed to a human reviewer.

Enable/Disable AI Item Review Toggle at Blueprint Level

A new "Enable AI Item Review" toggle has been added under Authoring > Blueprints > Settings > AI tab, allowing External Organization Admins, Internal Organization Admins, and SuperAdmins to control whether AI-powered item review is active for a given Blueprint.

When enabled, newly submitted items are automatically routed to Certiverse AI as the first review step; when disabled, the standard human review flow remains unchanged. New Blueprints default to ON, while all existing Blueprints are set to OFF upon release. An info tooltip on the toggle explains the feature's behavior.

Display AI Review Feedback in Revision View and Item History

AI reviewer responses are now provided to item writers in two contexts: 

  • In the revision view, when an item is returned after an AI review, only the answers and feedback for review questions where a problem was identified are displayed — consistent with how human reviewer feedback behaves.

  • In Item History, the full AI review record is visible, including answers and feedback for all review questions regardless of outcome. 

In both contexts, the reviewer is identified as Certiverse AI with a dedicated icon instead of a user's avatar.

Exam Management

Record AI Review Lifecycle Events in Item History

All AI review lifecycle events are now recorded in Item History, providing a complete audit trail consistent with human reviews. When an AI review starts, a "Review Started" entry is created attributed to Certiverse AI; when it completes, a "Review Completed" entry captures the outcome (Clearly Pass or Clearly Fail) along with the AI's answers and written feedback for each review question.

These entries use the same outcome labels and status model as human reviews, ensuring a unified history view. Technical failures of the AI service do not produce Item History entries.

Delivery

Extend Video Recording After Termination to Capture Termination Screen

The proctoring system now continues recording for up to 10 seconds after a termination incident occurs before closing the video stream, ensuring the termination message shown to the candidate is captured in the session playback. This provides clear evidence for dispute resolution and sponsor confidence that the candidate saw the termination reason. If the candidate manually closes the session before the delay elapses, the recording ends immediately. There is no impact on other termination workflows or system stability.