Practice6 min read
Deflection is a vanity metric. Measure resolution instead.
A ticket that never reached your team is not the same as a customer who got what they needed. Here is how we count the difference.
Most AI support tools report deflection: the share of conversations that never reached a person. It is an easy number to make go up. Hide the contact button, loop the customer through three articles, and deflection looks wonderful while your churn quietly grows.
We think the only honest number is resolution: did the customer get what they came for, and did they stay satisfied afterwards?
How Volt counts a resolution
A conversation counts as resolved when all three of these are true:
- Volt closed it without a human taking over.
- The customer did not reopen it, or open a new ticket on the same topic, within 72 hours.
- The customer did not leave a negative rating.
Anything else is either a handoff or an unresolved conversation. Neither is billed.
If a metric can improve while customers get angrier, it is the wrong metric.
The three numbers on our dashboard
| Metric | What it tells you | Healthy range |
|---|---|---|
| Resolution rate | Share of tickets Volt fully solved | 60 to 90% |
| Handoff quality | Share of handoffs the human solved without asking again | above 85% |
| Reopen rate | Share of resolved tickets that came back | below 4% |
Watch them together. A rising resolution rate with a rising reopen rate means the agent is closing tickets it should not.
What to do this week
- Pull last month's tickets and tag the ones that were reopened.
- Compare your current tool's "deflected" count to that list.
- Run a Volt simulation on the same month and compare.
The gap between those numbers is usually the most useful chart a support lead can bring to a planning meeting.