Voice analytics
Analytics is the application-wide picture: how many conversations were processed in a date range, how many escalated, and how your agents compare. Unlike Topics & Sentiment and Grading, it is always there — whichever outputs the application enables.
- The Voice IntELIEgence module visible in your left-hand navigation.
- This permission: Voice · App — VIEW.
- An application with conversations that have finished processing. Every figure comes from the date range you set, never from all time.
If Voice IntELIEgence isn't in your navigation, your role doesn't have visibility for it — ask a workspace administrator, or see Permissions and module visibility.
Steps
Select Analytics in the application’s sections — the last one above Settings.
Use the From and To date pickers. They default to the last 30 days, and everything below is recalculated from the conversations in that window.
A row of figures summarises the range. Which ones appear is decided by the service rather than fixed in the page, so expect the set to reflect what your application actually produces — Conversations is always among them.
Conversation volume plots one point per day for Processed and Escalated. By channel breaks the same conversations down by where they came from — Manual, Konnect or Unknown — not by telephone or web.
Agent performance lists Agent, Calls, Avg handle (mean call length), Containment (the share of calls that did not escalate), Avg grade and Avg sentiment, busiest agent first. Conversations with no identified agent are left out of this table only.
What success looks like
- A row of figures sits above the charts, including Conversations.
- Conversation volume shows a Processed and an Escalated series.
- Agent performance lists a row per identified agent, or reads No agent activity in this window.
What counts as an escalation
A conversation counts as escalated when the analysis flags it, or when its processing failed. That one rule drives the Escalated series and Containment — so a run of failures raises the escalation count on its own.
If something goes wrong
| Symptom | Likely cause | What to do |
|---|---|---|
| Avg grade shows an em-dash. | Grade was off for the conversations in range. | Enable it — see Grading. It applies to new conversations only. |
| Avg sentiment shows an em-dash. | Topics & sentiment was off for those conversations. | Enable it — see Topics and sentiment. Nothing already processed is re-analysed. |
| No agent activity in this window. | No conversation in range has an identified agent. | Check the Agent on a conversation's Transcript tab — see Managing personas. |
| Zeroes and em-dashes everywhere. | Nothing finished processing between From and To. | Widen the range, or see Browsing conversations. |
| The numbers moved between visits. | They are recalculated each time, and conversations were still processing. | Look again once processing has finished. |
Next
- Grading — the scores behind Avg grade.
- Topics and sentiment — the scores behind Avg sentiment.
- Voice application settings — where outputs are switched on.