Version 4.2

5 January, 2026

Enhancements

VA

AI Agent Dashboard Developments:
The AI Agent dashboard now delivers deeper insights into performance and efficiency with enhanced monitoring capabilities.
- AI Agent Dashboard with Prometheus Metrics:
  A new AI Agents dashboard tab provides comprehensive performance monitoring with real-time metrics, including total requests, error rates, response times, tool calls, token consumption, and tool performance analytics.
- Cache Savings Widget for AI Agent Dashboard:
  A new widget displays cache savings metrics on the AI Agent dashboard, helping teams track performance optimization.
Improvements for Hybrid LLM & NLU Usage:
The Virtual Agent node now supports Advanced Conversation Routing, which combines LLM intelligence with NLU-based intent matching. When enabled, the system uses AI to determine the most appropriate intent.
Ask LLM Node Response Format:
The Ask LLM node now supports structured output through JSON schema configuration, allowing users to define specific response formats.
Grace Period for Data Source Failure Detection:
Data sources now include a grace period before being marked as failed, improving reliability and reducing false positives.
Secure Secret Retrieval from Secret Store:
Secrets can now be retrieved directly from the secret store without being written to environment files or code, enhancing security and reducing exposure of sensitive credentials.
Daily Conversation Audio Recordings API:
A new API endpoint enables the retrieval of daily conversation audio recordings with automated processing and configurable retention policies, supporting compliance and operational review workflows.
Audit Logs External API:
A new external API endpoint provides access to detailed audit logs, enabling comprehensive activity monitoring.
Thai Language Support:
Thai language support has been added, enabling Virtual Agent deployment in Thailand and Thai-speaking regions.
Swahili Language Support:
Swahili language support has been added, enabling Virtual Agent deployment in East African markets and Swahili-speaking communities.

Improvements

VA

Increased Character Limit for AI Agent Additional Info:
The character limit for the AI Agent node's additional info field has been increased.
Additional Instructions for Guardrails:
Guardrails now support optional "Additional Instructions" fields for both input and output content filters, allowing administrators to define custom guardrails.
Failed Port removed when HTTP node is async
The failed port has been removed from HTTP nodes in async mode, streamlining flow design.
Investigation for Portal and VA Session Counts Inconsistency:
Session count discrepancies between the portal and Virtual Agent have been resolved, ensuring accurate reporting across the platform.
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.