fact_agent_conversations

Each row is one conversation handled in the agent workspace (Sales): who took it, how it was assigned, how long the customer waited, how it ended, and how the customer rated it. It replaces the legacy fact_conversations — now with full history (the legacy table kept a rolling 3 months) and richer assignment/transfer detail.

What questions does it answer?

How many conversations does each agent/team handle, and how do they resolve?
What is the first-response time and how does it vary by team or hour?
Which conversations were transferred, and how many times?
What CSAT do agents get?
How did the conversation arrive — from which flow, inbound or outbound?

Columns

Column	Type	Description
`agent_conversation_id`	Int32	Unique identifier of the conversation
`company_id`	Int32	Your company (filtered automatically)
`session_id`	Int64	The conversation flow that handed over to an agent (join `fact_treble_sessions`); `0` if none
`contact_id`	Int32	Treble contact; `0` if none
`created_at`	DateTime64	When the conversation entered the workspace
`finished_at`	DateTime64	When it was resolved; `NULL` if still open
`assigned_at`	DateTime64	When an agent was assigned; `NULL` if never assigned
`status`	String	Current status: `unassigned` (waiting in queue), `assigned` (with an agent), `finished`
`finish_type`	String	How it ended: `MANUAL` (closed by the agent) or `AUTO` (closed automatically); empty while open
`rating`	Int32	Customer satisfaction rating 1–5; `0` if not rated
`is_redirected`	Bool	`true` if the conversation was transferred at least once
`agent_id`	Int32	Agent currently/last responsible; `0` if unassigned
`agent_name`	String	That agent’s name
`tag_id`	Int32	Tag (queue/group) the conversation belongs to
`tag_name`	String	Tag name
`team_id`	Int32	Team owning the tag
`team_name`	String	Team name
`channel_id`	Int32	WhatsApp line
`direction`	String	`INBOUND` or `OUTBOUND` origin of the underlying flow; empty if there is no linked flow
`poll_id`	Int32	Flow that originated the conversation; `0` if none
`poll_name`	String	Flow name
`transfer_count`	UInt32	Number of transfers
`first_agent_message_at`	DateTime64	First message sent by an agent; `NULL` if none yet
`first_response_sec`	Int64	Seconds from the conversation’s creation to the first agent message; `NULL` if no agent message yet
`helpdesk_contact_id`	String	Contact id in your CRM, when integrated
`contact_wa_id`	String	Contact’s WhatsApp id
`transfer_waiting_time`	Int32	Configured limit (minutes) to wait for the contact’s CRM owner agent before assigning to the pool. Special values: `-1` = assign to owner immediately, `-2` = wait until the owner is online. `NULL` when not configured
`last_customer_interaction`	DateTime64	Last message from the customer
`assign_type`	String	How the agent got it: `AUTO` (assigned by the platform) or `MANUAL`
`assignment_strategy`	String	Per-flow assignment strategy: `lowest_first`, `specific`, `owner`, or `random`; empty when the company-level default applies
`language`	String	Conversation language: `ES`, `PT`, `EN`, `FR`, `DE`, `IT`
`channel`	String	Channel type (currently `WHATSAPP`)
`synced_at`	DateTime64	When this row was last written/corrected

Example queries

Volume and resolution per team, last 30 days

SELECT
    team_name,
    count()                                        AS conversations,
    countIf(status = 'finished')                   AS finished,
    round(avg(first_response_sec) / 60, 1)         AS avg_first_response_min,
    round(avgIf(rating, rating != 0), 2)           AS csat
FROM fact_agent_conversations
WHERE created_at >= today() - 30
GROUP BY team_name
ORDER BY conversations DESC

Conversations transferred more than once

SELECT agent_conversation_id, agent_name, transfer_count, created_at
FROM fact_agent_conversations
WHERE created_at >= today() - 7
  AND transfer_count > 1
ORDER BY transfer_count DESC

To see the full bot conversation that preceded the handover, join session_id against fact_treble_sessions and fact_treble_session_messages.

Get Started

Meta

Build with Treble

Main Platform

Agent Platform

Integrations

Analytics Warehouse

Data Warehouse (Legacy)

fact_agent_conversations

fact_agent_conversations

What questions does it answer?

Columns

Example queries

Volume and resolution per team, last 30 days

Conversations transferred more than once

​fact_agent_conversations

​What questions does it answer?

​Columns

​Example queries

​Volume and resolution per team, last 30 days

​Conversations transferred more than once

fact_agent_conversations

What questions does it answer?

Columns

Example queries

Volume and resolution per team, last 30 days

Conversations transferred more than once