AI Agent Action – Call Transcript Tool, Feedback Flow & Stability Improvements

Call Transcript Tool (Built-in)

The AI Agent action now features a built-in Call Transcript tool that allows the agent to access a contact’s call history. This tool operates internally with no configuration required; simply reference call transcripts in your agent prompt (e.g., “Use this contact’s call transcripts to summarise their concerns”), and the agent will automatically retrieve and utilise them.

This enhancement builds on the previously released conversation history tool, as the agent now has access to both past chat conversations and call transcripts, providing a more comprehensive view of every contact interaction.

Feedback Flow in Execution Logs

You can now submit structured feedback on each AI Agent run directly from the execution logs. Flag any issues you encounter, note what worked well, and share specific details for each run. All feedback contributes to improving the AI Agent’s accuracy, evaluation quality, and goal alignment over time.

Bug Fixes and Stability Improvements

A number of bugs have been resolved across the AI Agent action to enhance reliability and improve the overall experience.

The Call Transcript tool significantly increases the AI Agent’s contextual awareness. Rather than relying solely on chat history or static contact fields, it can now incorporate voice interactions, making its decisions, responses, and actions more grounded in the actual customer context. The feedback flow bridges the gap between the agent’s operations and your expectations, enabling faster iterations and minimising deviations.