|
Batch adjust start times in the schedule editor
Administrators and schedulers can now select multiple shifts or activities and batch edit their start times in the schedule editor. Users can move the selected items earlier or later by a specified duration, or set them to the same start time, without changing the length of the shifts or activities. This enhancement reduces manual effort and improves efficiency when making large-scale schedule updates.
Import decision metrics for workforce management
Workforce management administrators can now import and export decision metric performance and tiebreak values for use in automated allocation and approval processes. Administrators can upload or download CSV files containing agent values, eliminating the need to use REST APIs to import or update decision metrics.
Time-based filtering for opportunities and enrollments
Schedulers can now filter Opportunities and Enrolments by time as well as date. This enhancement enables users to narrow results to a specific time range, making it easier to find relevant records in high-volume environments.
|
Evaluation and survey forms division awareness
Administrators can now manage evaluation and survey forms by division. This enhancement helps ensure that users can view and modify only the forms available within their assigned divisions, improving organizational separation and access control. In some workflows, users may need to select a division when choosing forms.
Removal of transcription confidence threshold configuration
Genesys Cloud removes the transcription confidence threshold configuration by setting the default threshold to 0 and removing the setting from the Speech and Text Analytics Settings page. As transcription accuracy has continued to improve, confidence-based filtering provides less value than it previously did. All transcription segments are now preserved regardless of confidence score, helping ensure complete transcripts and improving downstream analytics such as sentiment analysis and topic spotting without requiring manual threshold tuning.
Segment an interaction transcript and drill down into specific areas of interest
Administrators and supervisors can now use generative AI to automatically segment interaction transcripts into meaningful sections, such as Greeting, Complaint, and Resolution. Similar to chapters in a book, these sections create a structured outline that helps users quickly understand the conversation flow and navigate directly to specific points of interest without reading the entire transcript.
|
|
Using custom attributes for knowledge filtering within Agent Copilot
Administrators can now configure Agent Copilot knowledge filtering to use conversation custom attributes as customer context, in addition to existing knowledge labels and categories. This change helps ensure that agents see knowledge articles that are relevant to the current interaction without requiring administrators to maintain separate knowledge structures for different contexts.
Agent Copilot enhanced conversational understanding for knowledge searches
Administrators can now enable enhanced conversational understanding for Agent Copilot knowledge searches. This feature helps Agent Copilot identify when a knowledge search is needed and use conversation context to improve search queries. Agents can also view the query used to retrieve knowledge suggestions. These enhancements improve the relevance of suggested knowledge, reduce unnecessary searches, and help agents find information more efficiently.
Intent health insights for Agent Copilot
Agent Copilot now provides intent health scores and utterance-level recommendations to help administrators evaluate and improve intent performance. Administrators can identify issues such as intent conflicts, utterances that are too short or too long, and outlier utterances that may negatively affect intent accuracy. Recommendations are prioritized by impact, enabling administrators to focus first on the utterances and intents most likely to improve training data quality and overall intent performance
|