Agent Copilot summarization model improvement

Administrators can now use the updated summarization model within Agent Copilot to improve support for outbound engagements and enhance currency recognition. The improved model helps improve the accuracy of summaries in scenarios that involve outbound communication, thus capturing key information more effectively. Also, the model now better identifies and accurately represents currency values, making financial interactions clearer and more precise. This enhancement builds upon the previous summarization capabilities to provide more robust and accurate assistance for agents.

Introducing Architect workitem flows for enhanced work automation

Flow authors can now define and manage workitem flows directly within Architect via dedicated actions tailored specifically for workitem processing. This feature enables administrators and contact center managers to improve operational efficiency by reducing the need for manual data actions to perform routine tasks, such as adjusting priority, changing status, or transferring workitems to ACD queues. By integrating these actions, Architect offers increased flexibility and extensibility for workflow designs that require collaboration across teams. For instance, a travel booking workflow can now validate incoming data, proactively request missing information through the customer's preferred communication channel, and route tasks to different teams such as insurance verification and claim processing.

Agents can now identify new and existing customers using non-E.164 phone numbers in both website and application interactions. With this update, non-E.164 numbers can be included as identifiers in web and app interactions, allowing agents to recognize and connect with customers who use local or alternative number formats. This enhancement supports smoother customer identification across diverse numbering formats, improving the experience for agents handling a broader range of interactions.

Supervisors can now use the agent status widget in performance dashboards to view all interactions handled by agents, along with the durations of the shortest and longest interactions. This feature allows supervisors to set an order preference based on interaction duration. Also, this feature allows supervisors to quickly access interaction details with a single click in the new dashboard UI. These enhancements help supervisors efficiently monitor agent performance in real time, providing a streamlined view of agent activity to support timely guidance and oversight.

Sentiment and agent empathy analysis is now available for German Switzerland (de-CH).

Administrators can now access three new conversation events in the Operational Console and they can also view the new events in the Event Catalog. This feature helps administrators gain early visibility into conversations that may become toxic in the Analytics service. The three event types that signal a toxic status are:

• Conversation-2000 - Too Many Events
• Conversation-2001 - Too Many Communications
• Conversation-2002 - Too Many Segments

Agent Copilot knowledge surfacing answer highlight and AI generated answers for manual search support is now available for Dutch Netherlands (nl-NL), French Canada (fr-CA), French France (fr-FR), German Germany (de-DE), Japanese Japan (jp-JP), Portuguese Brazil (pt-BR), Spanish Spain (es-ES), and Spanish United States (es-US).

Administrators and flow authors can now use the new search box within the Architect toolbox to quickly and easily locate specific actions. This feature alleviates the need to navigate through multiple categories to find the right action, which helps to improve efficiency when flow authors create and configure flows.

On October 7, 2024, Genesys announced that it would end support for Nuance Recognizer as a Service, Nuance TTS, and Nuance Mix Dialog Bot on that date. Genesys postponed that date to May 31, 2027.