From Bright Pattern Documentation

Revision as of 16:49, 22 June 2026 by Emma.stutzman (talk | contribs) (Updated via BpClonePage extension. Source page: draft:Agent-Desktop-2.0/QualityManagementGuide/ConductingEvaluations)

< Previous | Next >
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Conducting Evaluations

When an evaluator opens an interaction for review, the Eval Console provides the tools needed to review interaction content, complete evaluations, perform calibrations, and review evaluation history. Depending on permissions and available interaction assets, the Eval Console may include:

  • Evaluation forms
  • Interaction recordings
  • Chat transcripts
  • Screen recordings
  • Agent notes
  • AI-generated summaries and auto-scoring results

Evaluations and calibrations are both used to assess interaction quality. An evaluation is a review completed by an evaluator and assigned to an agent. A calibration is a shared review activity used to compare scoring consistency among evaluators.

Users with the appropriate quality management privileges can assign evaluations or calibrations to evaluators. To create assignments, select one or more interactions in the Eval Home search results and click Assign or Calibrate.

Evaluators are notified of new assignments through Agent Desktop alerts. Evaluators may also use saved searches to locate interactions and begin evaluations independently.

The following article describes the controls and sections used to conduct evaluations in Agent Desktop 2.0.


Evaluation Workspace in Agent Desktop 2.0


Evaluation Tabs

When an interaction is opened for review, the Eval Console displays a series of tabs across the top of the workspace. Use these tabs to navigate between interaction details, transcripts, recordings, evaluation forms, and evaluation history.

Available tabs may include:

  • Summary
  • Transcript
  • Screen Recording
  • Evaluation
  • All Evals
  • AI Custom Question
  • Autoscoring Form
  • Autosurvey Form
  • Real-time Checklist

Summary

The Summary tab is displayed by default when an interaction is opened for review. This tab displays high-level interaction information and participant activity.

Depending on the interaction type and enabled features, the Summary tab may display:

  • Interaction details
  • Participant information
  • Disposition data
  • Conversation statistics
  • Agent notes
  • Historical AI summaries


Summary tab


Participant-Based Navigation

Instead of navigating interaction segments with directional arrows, Eval Console displays a participant list within the Summary tab.

When available, selecting a participant from this list narrows the scope of the transcript and playback activity to interactions involving that participant.


Participant-based interaction navigation


A compact timeline control at the bottom of the screen provides a full interaction overview and allows users to switch between participant-focused playback and all-participant playback.


Timeline overview and participant playback controls


Please note that audio playback begins at the moment the selected participant joins the interaction. To hear IVR prompts or earlier conversation segments, switch to All Participants mode.

Transcript

Click the Transcript tab to view the interaction transcript.


Transcript tab


Transcript playback automatically synchronizes with the interaction timeline. By default, the transcript is limited to the portion of the interaction during which the selected participant was active. To view the complete interaction transcript, switch to All Participants mode.

Screen Recording

If screen recording is enabled for the interaction, the Screen Recording tab displays recorded desktop activity associated with the interaction.

Evaluation

Click the Evaluation tab to complete or review evaluation forms for the interaction.

Evaluation forms are assigned per service and may contain multiple areas and questions.


Evaluation tab


In Agent Desktop 2.0, evaluation forms are organized into collapsible sections with completion indicators.

This layout allows evaluators to more easily review progress and navigate long evaluation forms.

Form Selector

The form selector allows evaluators to select the appropriate evaluation form for the interaction.

Notes

Evaluators may attach notes to individual evaluation questions.

Notes are displayed on the interaction timeline and correspond to the playback position active at the time the note was created.

To create a note, hover over a question and click the note icon.


Click the note icon to create a note


Attachments

Evaluators may attach files to evaluation questions.

Attachments can provide supporting reference material for agents reviewing evaluations.

To add an attachment, hover over a question and click the attachment icon.


Click the attachment icon to add an attachment


Download

The Download arrow at the top of the screen downloads a ZIP archive containing available interaction assets.

Depending on interaction type and enabled features, the ZIP archive may contain:

  • Audio recordings
  • Chat transcripts
  • Evaluation spreadsheets


Download Option


All Evals

Click the All Evals tab to view evaluations and calibrations associated with the interaction.

Evaluations can be selected from a drop-down list.

Displayed items may include:

  • Draft evaluations
  • Pending evaluations
  • Accepted evaluations
  • Disputed evaluations
  • Completed calibrations

Confirmation

If evaluation confirmation is enabled, users may have the option to:

  • Accept evaluations
  • Reject evaluations

Rejected evaluations may include comments explaining the rejection reason.

Re-opened evaluations are retained for reference purposes but are not included in reporting calculations.

Intelligence Engine Features

If Intelligence Engine features are enabled for the contact center, evaluators may also have access to:

  • Autoscoring
  • AI-generated summaries
  • AI custom questions


Autoscoring tab may be visible if the Intelligence Engine is enabled
< Previous | Next >