Version 4.7

  • Published on Apr 3, 2026
  • 3 minute(s) read
  • eA
 Prev Next

Version 4.7

3 April, 2026

Enhancements

AI Agents

  • Knowledge Base Reporting:
    New Knowledge Base tab has been added under the Insights section, enabling users to monitor RAG performance directly within the platform.

  • Knowledge Base Hybrid Search:
    Knowledge Base now utilizes Hybrid Search by default, combining semantic and keyword search to deliver more accurate and reliable retrieval results.

  • Knowledge Base Filtering and Sorting:
    The Documents and Websites sections of the Knowledge Base now support column-level filtering and sorting by name, URL, and training status, with a Clear Filters option available when active filters are applied.

  • AI Insights:
    A new AI Insights feature is now available for projects on the Premium edition. When enabled, Summary, Topic, Churn Risk, Conversation Result, Non-Resolution Risk, Actions, and Improvement Areas are automatically generated at the end of each session and made available as system variables in Post Session Flow. Insight types and custom descriptions can be configured per project through the dedicated AI Insights tab in Project Settings.

  • AI Testing LLM Settings:
    A dedicated AI Testing tab has been added to Tenant Settings for configuring the LLM model used during evaluation.

  • AI Testing - Evaluation Integration:
    AI Testing simulations are now evaluated using the evaluation endpoint, providing automated performance scoring for each simulated conversation based on the success criteria defined in the test configuration.

  • AI Testing Evaluations - Widget:
    A new AI Agent widget has been added to the AI Testing Evaluations dashboard, displaying average input token consumption, average output token consumption, and average response time for AI Agent calls across simulation conversations.

  • AI Agent Dashboard – Token Usage Widgets:
    Two new widgets have been added to the AI Agent dashboard: Average Tokens per Session and Average Tokens per Minute. Both widgets display Input, Cached Input, and Output token breakdowns along with total token counts.

  • AI Agent Invoke Count Tracking:
    The total number of AI Agent invocations per session is now recorded in the Sessions table, enabling more accurate usage and cost analysis.

  • Webhook Node Timeout:
    A configurable timeout option has been added to the Webhook Listener node, preventing sessions from getting permanently stuck when an expected callback is not delivered.

  • SessionChatHistory Variable:
    A new static variable, SessionChatHistory, is now available across all flows. It captures the full conversation history up to the point it is used, enabling scenarios such as generating summaries during live agent handovers.

Improvements

AI Agents

  • Redis Cluster Support: The platform now supports Redis Cluster configuration, improving scalability and resilience for high-availability deployments.

  • RabbitMQ Queue Reliability: RabbitMQ queue configuration has been updated to improve message reliability and fault tolerance, reducing the risk of message loss in high-availability deployments.

  • lastAiAgentName Context Behavior Update: The lastAiAgentName static context variable now stores the name of the most recently responding AI Agent, enabling flow designers to branch logic based on which agent produced the last response.

  • TestChat VAD Settings:
    Default, minimum, and maximum values have been defined for all VAD parameters in TestChat Settings, providing clearer boundaries during voice testing.

  • ElevenLabs TTS Integration:
    The ElevenLabs TTS configuration interface has been updated with a new model option and a revised placeholder message for the Custom Model field.

  • AI Testing UI Revisions:
    Several usability improvements have been made to the AI Testing interface, improving the overall test creation and evaluation experience.

  • Knowledge Retrieval Tool Query Customization:
    The query parameter description in the Knowledge Retrieval tool can now be edited by users, allowing custom descriptions to be defined for project-specific retrieval requirements.

  • 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.