Analytics Suite

At the highest level, analytics show how users move through the conversational funnel — from entry into a session through to completion. This helps identify points of friction where users disengage or abandon the process.

Filter by agent version

What it is: You can now filter analytics by a specific agent version or view all versions together.
Why it matters: Compare performance across versions, validate changes, and identify what improved or regressed.
How to use: Open Analytics. Use the Version selector in the top-right corner to choose a specific version or All versions to view project-wide performance.

Core Metrics

Conversational Funnel – visualises the conversational throughput of all sessions (how many clients started the interaction and when they left).
Session Length – total and average duration of conversations, showing the depth of engagement.
Bounce Rate – proportion of sessions abandoned before any meaningful interaction.
Relational Metrics – qualitative measures such as turn balance, pacing, or user sentiment.
Global vs. Per-Agent Views – compare aggregate performance across a project with the results of individual agents.
Multimodal Activity – counts of structured actions such as forms filled, images displayed, or web views opened.
Memory Usage – tracks how often memories and Knowledge Graph connections are accessed by agents and their impact on personalization or resolution times.
Conversion Summaries – high-level indicators of how conversations connect to business outcomes (resolved cases, completed workflows, influenced sales).

These metrics allow product teams to continuously improve agents: reducing drop-off, shortening resolution times, and optimising relational quality. They also provide evidence for business leaders on how relational intelligence contributes directly to KPIs like customer satisfaction, efficiency, and conversion.