Skip to main content
This document provides information about the latest feature updates and enhancements introduced in Quality AI of AI for Service (XO) v11.x releases. For previous updates, see release notes of 2025.

v11.23.1 April 11, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Analyze

Custom Fields Support in Conversation Mining, Audit, and Reports The platform now retains business-specific custom fields ingested via Express File (CSV columns or push API fields) and Agent AI integrations (custom data key-value pairs). Supervisors, QA teams, and API consumers can filter, analyze, and export conversations based on these fields — no prior configuration required. Learn more→

Configure

Direction-Based Evaluation and Reporting for AutoQA and Conversation Intelligence Quality AI now supports contact direction (Inbound and Outbound) as a dimension across AutoQA, Conversation Intelligence, dashboards, reports, APIs, and Conversation Mining. Supervisors can configure separate evaluation forms and scorecards by direction and channel at the queue level, ensuring conversations are assessed against criteria that match their operational context. Learn more→

v11.23.0 March 28, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Analyze and Configure

Minimum Duration Threshold for AutoQA and Scorecards You can now set a minimum interaction duration threshold in evaluation forms and agent scorecards to ensure that AutoQA evaluates only the meaningful conversations. Before scoring, the system checks each interaction’s duration and excludes short or incomplete ones from quality metrics. Contacts excluded from scoring and quality calculations remain visible. Learn more→

v11.22.1 March 14, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Role Management

Update of Quality AI Permissions for Default Roles App Developers and App Testers can now access Quality AI through their default roles, without needing custom roles. This update revises default role permissions to give developers and testers appropriate access while maintaining the right level of control. Learn more→

Configure

Manual Evaluation Metric A new Manual Evaluation metric type is now available for QA-only assessment of complex and nuanced scenarios. This metric is supported only in points-based evaluation forms and is excluded from AutoQA, Agent Attributes, and Agent Scorecards. Manual metrics are clearly labeled in reports and APIs, and unaudited conversations show no AutoQA response for these metrics. Dynamic By Question (Speaker-Based Adherence) The Dynamic By Question metric now supports speaker-based answer adherence. Admins can configure the answer detection speaker — Agent or Customer — based on trigger rules. When the trigger speaker is an Agent, an optional scoring setting enables sub-weightages and partial scoring for both trigger and answer adherence, and auditors can manually evaluate both in the Audit Screen. Conversation Mining now includes a Not Applicable filter for Dynamic metrics, and reporting and heatmap logic treat trigger absence as Not Adhered when the scoring option is enabled.

v11.22.0 February 28, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Configure

Configurable Crosstalk Evaluation for By Speech Metrics The Crosstalk metric now detects simultaneous speech between the agent and customer, including customer interruptions. The Dynamic By Question metric supports speaker selection and sub-weight assignment for agent-triggered adherence, and answer detection can now be extended beyond the agent — enabling use cases such as customer confirmation and verification. Points-Based Scoring for Complex Evaluation Forms Evaluation forms can now use points-based scoring, making it easier to build and manage complex forms with more than 20 metrics. QAs can assign weights by points rather than percentages, and all points-based forms include audit tracking for score changes and a full record of updates. GenAI Logs Enhancement in Audit Screen The Audit Screen now displays detailed GenAI call logs at the conversation level for easier debugging. Logs are organized by GenAI feature in expandable dropdowns that show only enabled features, and can be filtered by Success or Failure status. Each log entry includes date and time, GenAI feature name, language, model name, integration type, prompt name, token usage, response duration, and full request and response payloads.

v11.21.1 January 31, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Analytics

AI Justifications for Gen AI Question Metrics Extended to Reports and APIs AI Justifications for Gen AI by Question metrics are now available in the Interaction Evaluations and Conversation Analytics reports, and through APIs. When enabled, the system provides AI-generated explanations for each evaluation score, extending this capability beyond the existing UI.

Configure

Enhanced Taxonomy Builder, Topic Discovery, and Resolution Detection Taxonomy Builder and Topic Discovery now offer improved usability with clearer visual hierarchy, contextual tooltips, sentiment and resolution-based bubble coloring, and enhanced filtering and navigation. Resolution Detection is now configurable at the app level. You can choose between topic-based detection for strict matching or LLM-based assessment for holistic evaluation. This flexibility helps you accurately classify interaction outcomes based on whether primary issues are resolved.

Analyze

Conversation Intelligence Dashboard Updates The Conversation Intelligence dashboard is now split into two specialized dashboards:
  • CX Insights introduces new widgets to help you understand customer experience, including Resolution Rate tracking and CSAT and DSAT Drivers powered by driver-impact scoring with detailed warnings.
  • Performance Insights enhances agent monitoring with a trendline for the Kore Evaluation Score.
Legacy widgets, including Contact Center Efficiency, Agent Occupancy, and Agent State Monitor, are now deprecated but remain available via feature flag for existing users.

v11.21.0 January 17, 2026

This update includes enhancements and bug fixes. The key enhancements included in this release are summarized below.

Analytics

Evaluation Form Summary Report Enhancements The Evaluation Form Summary report now includes a Total Interactions column to show the overall interaction count. The Total Applicable Interactions column excludes inapplicable interactions for dynamic-by-question metrics where the trigger was absent. These enhancements improve calculation accuracy and ensure that totals and percentages align with Heatmap data.

Analyze

Audit Allocation Enhancements Audit Allocation management uses a dedicated “Allocations” menu and a “My Allocations” tab. QA managers can track auditor progress, create and edit custom allocations, assign interactions to agents by percentage or count, and reassign pending interactions to manage availability while maintaining consistent quality coverage.

Configure

Agent Queue Management in the Quality AI UI Users can assign agents to Agent AI and Quality AI Express queues directly in the Quality AI UI, eliminating the need for public APIs. This update supports platform-level users, provides visibility into queue IDs for API use, and keeps agent mappings entirely separated from Contact Center AI (CCAI) routing and configuration. SFTP Chat Script Timestamp Enhancements Quality AI Express now supports offset-based timestamps for chat script ingestion via Secure File Transfer Protocol (SFTP). With this update, users can configure the chat script timestamp format at the app level under Conversation Sources. When users select offset-based timestamps, chat script can include message-level offsets relative to the conversation timeline without start or end time validation during ingestion. The system uses start and end dates solely for reporting and filtering. This change applies only to chat conversation ingestion and doesn’t affect voice conversation ingestion.