Who this is for. AP Helpdesk Analysts and Managers.
Availability. This panel appears only where the instance runs in collaborative mode. See Settings.
This panel reports how the queue is being worked rather than what is in it: who is handling tasks, how they were answered, and how much mail needed an ad hoc look.
The panel carries its own range selector, offering 1 Day, 7 Days, 30 Days, and 365 Days, and prints the dates it covers beside it. Ranges are set per panel, so check the selector before comparing one panel against another.
A Filter control above the panel switches the figures between Team and User, so the same measures can be read for the desk as a whole or for one person.
Task handling
Five figures describing what happened to the tasks in the range.
Table 1. Task handling measures.
Measure |
What it counts |
|---|---|
Open |
Tasks still being worked, with their share of the total. |
Done |
Tasks completed, with their share. |
AI Agent Replies Sent |
Replies the agent sent. |
User Replies Sent |
Replies a person sent. |
No Reply Sent |
Tasks completed without anything going back to the sender. |
The two reply counts are the measure of what collaborative mode is costing and buying. A high user share means people are writing replies the agent prepared; a high No Reply Sent count means tasks are being closed without an answer, which is worth understanding rather than assuming.
Ad hoc email review
Three counts of mail that needed a person to look at it despite not resolving to a supplier contact, by sender type: External Cohort Sender, Internal Cohort Sender, and Unknown Sender.
This is the panel that shows what poor contact data costs in attention. Every one of these is a message the agent could not attribute confidently. See SOR Insights.
Tasks by detected intent
Beneath the status figures, Tasks By Detected Intents breaks the same tasks down by what the agent recognised. It is a table rather than a chart, and it carries timing columns the other dashboard panels do not.
Table 2. Columns in Tasks By Detected Intents.
Column |
What it reports |
|---|---|
Detected Intents |
The intent, including No AP Business Intent for tasks the agent could not place, and the macro intents Request, Inquiry, Update, Acknowledgement, Unclear, and Unintelligible. |
Count and Percentage |
How many tasks carried that intent, and its share. |
Max Age and Average Age |
How long tasks of that intent have been waiting, as the worst case and the mean. |
Max Resolution and Average Resolution |
How long they took to complete, once complete. |
Read age against resolution. A long average age with a short average resolution means the work is quick once someone picks it up, so the delay is in the queue rather than in the task. Intents with no activity in the range report zero rather than being hidden, so the table lists every intent your agent handles.
Figure 1. The Tasks and Utilization panel, filtered by team.