Conversation Toolkit 2

Conversation Toolkit 2 is an AI-powered analytics add-on designed to analyze chat and voice conversations and extract insights like Sentiment, Clarity, Friction, Outcome and more. It processes conversations both in real time and historically, automatically extracting a predefined set of structured metrics using the preconfigure AI Agent.

The extracted metrics are defined in an Admin workspace and can be adapted to different business scenarios, specific to each industry and process, including human – AI conversations, human to human conversations (live chat). This ensures consistent and scalable evaluation of conversations across channels and use cases.

When a virtual assistant also offers a Live Chat feature and conversations are handed over to human agents, Conversation Toolkit 2 becomes especially valuable for Contact Center and Customer Support operations. In this context, the solution enables organizations to better understand customer interactions and objectively evaluate agent performance by analyzing real conversations. Metrics can be used to assess aspects such as interaction quality, customer sentiment, resolution effectiveness, and adherence to communication standards.

By combining automated conversation analysis with alert-driven processing, Conversation Toolkit 2 transforms raw conversational data into actionable insights, supporting continuous improvement for both AI Agent behavior and human agent performance—without requiring manual reviews or sampling.

What’s New in Version 9.15.1

  • AI Analyzer — Automatic Outcome Generation. The AI Analyzer has been enhanced with the ability to automatically generate new outcomes and add them to the Outcome Dictionary when an analyzed conversation cannot be matched to any of the predefined outcomes.

    • Previously, if a conversation did not match any existing outcome in the Outcome Dictionary, it would remain uncategorized (no outcome). With this enhancement, the AI Analyzer can now recognize such gaps and propose new outcomes on the fly, enriching your Outcome Dictionary over time based on real conversation data.

    NOTE: This functionality is disabled by default. Solution administrators can enable it after importing the solution settings in the Conversation Toolkit Admin workspace. For more information, see Step 3. Import default settings.

Key Capabilities

  • Real-time and historical analysis of chat and voice conversations, human-AI and live chat conversations.
  • AI-driven extraction of predefined metrics (quality, performance, compliance, sentiment, sentiment, friction, etc.).
  • Predefined and custom metric configuration via an Admin workspace.
  • Structured metric storage for reporting and dashboards.
  • Complex reporting dashboards with charts and data views, enabling detailed analysis of conversation trends, agent performance, and metric evolution over time.

Key Components

The main components of Conversation Toolkit 2 include:

  • Druid Data Service back-end. Handles conversation storage, metric definition and values, conversation outcomes and processing states.
  • Two dedicated workspaces:
    • Conversation Toolkit Admin workspace – Used by solution administrators to:
      • Define and manage metric definitions, formats and thresholds.
      • Set specific business goals (outcomes).
      • Provision default settings.
    • Conversation Toolkit workspace – Used by business users to:
      • Review analyzed conversations.
      • Inspect extracted metric values.
      • Track processing status.
      • Reset and reassign conversations for analysis.
  • AI-driven alert flows. Automatically process conversations in real time or at predefined intervals.