Skip to main content
Open Quality & Testing → CSAT & Outcomes to connect customer feedback and resolution data back to AI Assistant work.

​
Reporting period

The current view supports:
  • 7 days;
  • 14 days;
  • 30 days;
  • 60 days; and
  • 90 days.
Changing the period updates both the summary metrics and ticket outcome list.

​
Metrics

The current outcome cards include:

​
CSAT Score

The percentage of rated tickets recorded as good. If no tickets were rated in the selected period, RAUM shows no score rather than pretending an empty data set is 100% satisfaction.

​
Good Ratings

Count of good Zendesk satisfaction ratings in the selected period.

​
Bad Ratings

Count and share of bad ratings among rated tickets. Bad ratings are useful inputs for AI Coaching, where poorly rated Assistant work can be reviewed in detail.

​
Average Resolution

Average recorded resolution time for resolved tickets, together with the number of resolved outcomes included.

​
Ticket Outcomes table

The table can show:
  • Zendesk ticket ID;
  • CSAT value;
  • satisfaction comment;
  • resolved state;
  • resolution time; and
  • outcome date.
Use ticket-level outcomes when a summary metric moves unexpectedly. Aggregate percentages are good at announcing a problem and famously poor at explaining one.

​
Data dependency

Outcome tracking depends on Zendesk events reaching RAUM. If the page remains empty even though Zendesk has ratings, verify the configured satisfaction-rating trigger/webhook under Zendesk → Automation.

​
Use outcomes with other evidence

Do not evaluate Assistant quality from CSAT alone. Pair outcomes with:
  • Answer quality on the dashboard;
  • Ticket Activity;
  • Coaching feedback;
  • Knowledge Gaps;
  • autonomy/handoff metrics; and
  • the actual customer conversation.
A resolved ticket with no complaint is not automatically a correct answer, and a bad rating can reflect factors outside the AI response.

AI Coaching

Investigate poorly rated Assistant tickets and propose reviewed fixes.

Zendesk Automation

Verify the satisfaction-rating trigger and webhook flow.