Supervisor — Quality Assurance Scenarios Hub

A page for supervisors controlling quality checks. Here you will conduct reviews, use checklists, and assess operator work without unnecessary navigation. Below are instructions for checks, checklists, and viewing assessments.

Quick Overview

• High-Level Goals and Scenarios
• Where to Work in the Interface
• Quick Links

High-Level Goals and Scenarios

Quality Checks

Goal — assess customer service quality and identify areas for improvement.

• Create assessment — how to create a new dialogue assessment with search or from a specific chat. → ./how-to/create-assessment.md
• Conduct quality check — how to assess a request using a checklist. → ./how-to/conduct-quality-check.md
• Run a bulk assessment from the chats table or a list of IDs — with the auto-assessment settings enabled: checklists by topic, repeat assessment, skip rules. → ./how-to/run-bulk-assessment-from-chats-table.md
• Share a chat selection with a link — how to create a protected link to a checked chat list and open it. → ./how-to/share-bulk-selection-link.md
• View the bulk assessment history — how to see the result of every run, the skip reasons and add a comment. → ./how-to/view-bulk-assessment-history.md
• View assessment details — how to see complete information about a conducted check and adjust AI assessments. → ./how-to/view-assessment-details.md
• View assessments list — how to find the needed assessment through quick search and filters. → ./how-to/view-assessments-list.md

Working with Checklists

Goal — use standardized checklists for objective assessment.

• Use checklist — how to apply a checklist for quality check. → ./how-to/use-checklist.md

Assessments

Goal — record check results and provide feedback to operators.

• Assess request — how to assign a score by various criteria. → ./how-to/assess-request.md
• Adjust AI assessments — how to correct AI errors and add comments to criteria. → ./how-to/adjust-ai-assessment.md
• Find out why the AI took points — how to get a verdict (AI, criterion, knowledge base or human) with quotes and a draft fix from the AI Assistant. → ./how-to/diagnose-ai-assessment-with-assistant.md
• Ask the AI judge — how to find out through the AI Assistant why AI passed or failed a criterion, without changing the assessment. → Ask the AI judge about a verdict
• Approve assessment — how to complete the check and make the assessment available to the operator. → ./how-to/approve-assessment.md

Where to Work in the Interface

Main work with quality happens in the Quality Assurance module:

• Quality Assurance module: Apps -> Quality Assurance — here you conduct checks and assessments.
• Dashboard: for viewing quality statistics.
• Checks: for conducting new assessments.

Quick Links

Checks

• Create assessment
• Conduct quality check
• Run a bulk assessment from the chats table
• View the bulk assessment history
• View assessments list
• View assessment details

Checklists

• Use checklist

Assessments

• Assess request
• Adjust AI assessments
• Diagnose an AI assessment with the AI Assistant
• Ask the AI judge about a verdict
• Approve assessment

Export and Analysis

• Export assessments
• View dashboard