Analytics

The Analytics page is the single source of truth for how your agents (voice and chat) are performing. Everything for the selected period lives on one scrollable page so you can flick from volume, to sentiment, to per-agent performance, to top contacts without losing your place, then keep scrolling into the voice-only deep dive at the bottom.

Live wallboards for the floor

Analytics is the analytical, slice-and-dice view. If you want the same data on a TV in the contact centre, build a Wallboard; there's a shortcut button in the header.

The Period Picker

A single period picker in the page header drives every chart on screen. Switch between:

  • Last 7 days: recent operational view.
  • Last 30 days: the default; month-style baseline.
  • Last 90 days: quarterly view.
  • This month: current calendar month, useful for live stand-ups.
  • Last month: closed comparison window.

Every section recalculates against the chosen period when you switch, including the previous-period comparison arrows that appear on most KPI tiles.

Omnichannel Summary

The first row of KPI tiles spans every channel: voice calls, web chats, and any text channels you have wired up, so you can answer “how busy were we overall this week?” in one glance. Each tile shows the headline number for the period plus a delta against the previous matching window.

Channel Volume & Channel Mix

Two side-by-side charts:

  • Channel Volume: a stacked time-series of conversations per channel, so you can see how voice, chat and the rest move together over the period.
  • Channel Mix: a compact donut showing the share of the period taken by each channel. Useful for tracking the long-term shift from voice to omnichannel.

Sentiment

Sentiment over time, computed from the conversation summary the AI produces at the end of every voice call and chat. Each point is bucketed into positive, neutral, or negative; the chart shows the daily mix so you can spot the day a change to a system prompt accidentally turned tone sour, or the day your refund policy went live.

Agent Performance

A side-by-side comparison of every agent in the org for the chosen period. For each agent you get total conversations, average duration / message count, and the per-agent success / completion rate. Sort the rows by any column to identify which agent is carrying the load and which one needs prompt or tool work.

Top Contacts

The contacts that interacted with your agents most often in the period, with their last conversation, channel, and dominant intent where the AI was able to detect one. Click a row to drop straight into the call log filtered to that caller, exactly the same flow as the Repeat Callers report, just rendered as a quick reference here.

Voice Channel Section

Below the omnichannel summary there's a clearly labelled Voice channel divider, after which the page drops into voice-specific charts. These come from the original analytics product and remain the deepest view of what your telephony stack is doing.

KPI Cards

  • Total Calls: calls handled in the period.
  • Total Duration: combined call time for the period.
  • Average Duration: mean call length.

Each card shows a growth indicator versus the previous matching period so you can see at a glance whether volume or efficiency is trending up or down.

Call Volume

A daily time-series of voice calls with the previous-period line drawn behind for comparison. This is the chart most people screenshot for weekly reviews.

Quality

The quality panel surfaces:

  • Success Rate: share of calls that completed without an error.
  • Average First Response Time: how quickly the agent starts speaking after the caller connects. Lower is better.
  • Error Rate: share of calls that hit a runtime error.
  • Service Level Score: composite score factoring in response time, completion rate, and error rate.

A sortable table of voice activity bucketed by month. Each row shows total calls, average duration, total duration, and completion rate. Click any column header to sort.

Busiest Hours

An hour-by-hour bar chart for the period, colour-coded by time of day (night / morning / afternoon / evening). Useful for spotting the peak hour to staff a human team around, or to schedule callbacks into a quieter window.

Regular monitoring

Review your analytics at least weekly. Consistent monitoring helps you spot issues early (a rising error rate or declining completion rate) before they impact your callers' experience. The Sentiment and Quality sections are the two to keep on your standing weekly agenda.