Call Analytics: Sentiment on Your Calls
Review AI sentiment by period, extension and number, while distinguishing analyzed channels from unique calls and audio quality.
Analytics → Call Analytics presents automatic sentiment analysis for call channels. Positive, neutral and negative labels help identify conversations to review; they do not directly measure customer satisfaction or prove how well a person handled a call.
Set the report
- Confirm the selected business/location in Office.
- Open Call Analytics and set From date and To date.
- Choose an Extension, or keep All extensions.
- Click Search and wait for the results. Use Reset to clear the filters.
Understand calls and analyzed channels
Total Calls and Analyzed Call Channels use different counting units. A call may have more than one channel, so the analyzed-channel count can be higher than the call count. Do not divide these two figures to calculate the percentage of unique calls analyzed.
The Positive, Neutral and Negative Sentiment indicators describe the analysis represented in the report. They are not counts of unique customers or manual satisfaction-survey responses. A missing category or unavailable value should not automatically be interpreted as a positive outcome.
Read the charts and table
- Sentiment Overall: review the positive, neutral and negative distribution for the displayed context.
- Sentiment by Quarter: review the quarterly labels and extension context shown on the chart when examining changes over time.
- Sentiment by Number: compare Local Number, Qty, Positive, Neutral, Negative, Sentiment Overall and Categories. Entries can identify extensions as well as business numbers; inspect the label on the row.
Use Search list and the 10, 25 or 50 rows selector to find entries. The table provides Copy, Excel, CSV, PDF and Print. Verify the result and any active table search before sharing an export.
Keep three different measures separate
- Sentiment: an AI interpretation of the conversation, which may contain errors.
- Audio quality: the call-quality information available in Call info and the Audio quality filter in All Calls.
- Call review: a person's manual star rating and note for a call.
A technically clear call may contain an unhappy customer, and a negative topic may be handled well. Review the recording or available transcript and the surrounding context before drawing conclusions. Do not treat a sentiment label as a final evaluation of a staff member.
If results seem incomplete or unexpected
Check the date range, extension and selected location. Compare the same context across reports and inspect the relevant individual calls. Fewer or more analyzed channels than calls is not, by itself, evidence of missing data. For unexpected labels, note the Call ID, period and extension when contacting support.
For AI consumption, use AI Usage and the AI Wallet guide. The sentiment chart itself does not establish the charge for an individual analysis.