Conversations by Disposition — Dedicated Sub-View

The Conversations by Disposition sub-view provides an expanded, full-page version of the disposition chart with the complete filter panel visible. It allows analysts to isolate specific disposition types, channels, or assistants to understand the breakdown of how conversations are ending.

Figure 1: Conversations by Disposition — Full Sub-View with Filter Panel (03/08/2026 – 04/09/2026)

1. Filter Panel

Filter	Description
Date Selection Type	Custom Dates (default). Supports relative options: Current Day, Current Week, Current Month, Current Quarter, Previous Day/Week/Month/Quarter.
Start Date / End Date	Shown as 03/08/2026 12:00:00 AM – 04/09/2026 12:00:00 AM in this example.
Channel	(All) — includes chat, phone, and any other configured channels.
Stage	STG (Staging environment). Options include PROD, DEV, STG. Always verify Stage = PROD for production analysis.
Assistant	(All) — or select a specific assistant to isolate its disposition pattern.
Disposition	(Multiple values) — allows filtering to specific disposition types for focused analysis.
Business Day	(All) — can be filtered to isolate business day vs. weekend performance.

2. Disposition Types — Definitions

Connection Disposition	What It Means
AppDisconnect (orange)	The application (platform or bot) terminated the conversation. This can indicate session management logic, bot-initiated endings after goal completion, or technical errors causing forced disconnection.
UserDisconnect (blue)	The user ended the conversation themselves by closing the chat, hanging up, or navigating away. Can indicate successful self-resolution or user frustration.
Transfer (green)	The conversation was routed to another destination (live agent, different flow, or external system). Presence of Transfer in this chart confirms a transfer routing occurred.
Timeout (not visible in expanded view)	The conversation ended because no user activity was detected within the configured timeout window. Common in voice channels when a caller does not respond.

3. Business Value

Understanding disposition distribution is critical for AI assistant quality assessment. A high AppDisconnect rate may indicate the assistant is terminating conversations prematurely. A high Timeout rate suggests users are becoming unresponsive — potentially because they are confused, frustrated, or the assistant is not prompting effectively. UserDisconnect is generally the most desirable outcome, as it implies the user completed their interaction voluntarily. The Transfer disposition, when present, directly correlates with the containment metrics tracked in the Containment tab.