Version 3.2

Version 3.2

29 May, 2025

VA

Improvements

VA

• Agentic AI:

  • AI Agent Guardrails:
    Guardrails for AI Agent has been introduced which allows users to enable key safety measures such as avoiding harmful content, staying grounded to verified information, detecting jailbreaks, and maintaining output consistency. Guardrail options can be selected during the creation step and modified from the AI Agent Details page.

  • User-Based Remembering of LLM Configurations:
    The system now remembers the last selected LLM configuration for AI Agent creation, editing, and import processes. This eliminates the need to re-enter the same settings repeatedly, streamlining the workflow and improving user experience.

• Integration with gen-ai Module:
  Virtual Agent is now integrated with the GenAI module via a proxy service, eliminating the need for separate integrations with individual providers. This allows the same task to be executed across multiple providers through a single endpoint, streamlining development and increasing flexibility. The integration supports the following features:

  • Utterance Generation
  • Summarization
  • Text Response Generation
  • Generate Project with AI
  • Generative Coding Assistant

• Adding Document Filtering to GenQ&A:
  GenQ&A users can now enable a new Document Filtering toggle in the node’s advanced settings. This feature ensures answers are sourced from a specific document, improving response reliability, consistency with internal knowledge, and reducing irrelevant information.

• Project Settings Versioning:
  Project settings including graceful publish, spell correction and real-time translation features are included in the project versioning.

• Enabling Retention as an integer value:
  The retention feature in tenant settings now supports custom integer values, allowing users to define a specific retention period. The allowed range is from 1 (minimum) to 120 (maximum).

• Dashboard & Conversation Pages Filter Changes:
  Several updates have been made to improve responsiveness and user experience on the Dashboard and Conversations pages:

• Magpie & Raven Memory Leak Issues:
  Memory usage in Magpie and Raven that shows potential memory leak issues has been resolved.

• Reducing Call Dispatch Interval in Outbound:
  The interval for sending all calls from the list to Hummingbird IVR has been reduced from 5 minutes to 2 minutes. This change speeds up the triggering process and improves overall system responsiveness.

• Minor UI Improvements:
  Several minor improvements have been made to the user interface to enhance user experience and usability.

• Minor Bug Fixes:
  Various minor bugs have been addressed to improve overall system stability and performance.

LC

• Calculating Waiting time with message:
  Customers are now informed with a system message showing their estimated wait time and queue position when reaching the Live Chat node. This helps manage expectations, increases trust in the system, and reduces uncertainty during the handover process.