The five KPI stat cards at the top of the Home Dashboard give you a quick read on the health of your support operation. Each one shows a single number for the selected period.
First response time
What it measures: The median time between a conversation being created and the first reply sent by an agent.
A lower number is better. This is one of the most important metrics for customer experience — customers who wait too long for a first reply are more likely to leave negative feedback or abandon their request.
What affects it:
- How quickly agents pick up new conversations
- Whether conversations are sitting unassigned
- Business hours and team availability
Median first response time
Median first response time is the middle value of how long customers waited for a first reply from a team member, so a few unusually slow or fast tickets don't skew it.
The clock starts when the conversation is created. If the AI agent handled it first and then handed it to your team, the clock starts at the handoff instead — so time the AI spent on it isn't counted against your team. It stops at the first human reply.
Spam conversations are excluded. If you turn on the business hours option, the time is measured only within your configured business hours.
Resolution time
What it measures: The median time from a conversation being opened to it being resolved.
A lower number is better. This measures how efficiently your team closes conversations end to end, not just how fast they respond.
What affects it:
- Complexity of customer issues
- Number of back-and-forth messages needed
- Whether conversations are getting stuck or forgotten
Median resolution time
Median resolution time is the middle value of how long it took to resolve conversations, measured from when the conversation was created to when it was resolved.
Conversations are counted in the period they were resolved, not the period they were created. Spam is excluded. If you turn on the business hours option, only time within your configured business hours is counted.
Messages per ticket
What it measures: The average number of messages exchanged before a conversation is resolved.
A lower number is generally better — fewer messages means the issue was resolved more efficiently. Very high numbers may indicate unclear communication, missing information from customers, or issues that need better guidance articles.
Escalation rate
What it measures: The share of conversations that were handed from the AI agent or a first agent to someone else.
A lower number is better. A high escalation rate may mean the AI agent is struggling with certain topics, or that agents are passing conversations on too readily.
What affects it:
- AI agent guidance coverage
- Team structure and assignment rules
- Conversation complexity
One touch rate
What it measures: The share of conversations resolved with a single agent reply.
A higher number is better. A high one touch rate means your team is resolving issues in one go without multiple back-and-forth exchanges — a sign of efficient, clear communication and good self-service content.