Analytics
Figures you can trace to the records behind them
Every number on a dashboard, in a report or from the API comes from one metric dictionary, and an Explain button shows the records that produced it.
What you can see
Live dashboards
Waiting, offered and assigned work, the longest wait and agents by state, by channel, queue or agent, updated as it happens.
Drill-downs
History in 15-minute to daily buckets, and the conversations behind any figure, one click away.
Customer journeys
Channel paths, repeat contacts within 7 days and handovers from the AI agent to people.
Saved and scheduled reports
Reports by channel, queue, agent or campaign, e-mailed as CSV daily, weekly or monthly.
Threshold alerts
Queue wait, abandon rate, service level and adherence rules that alert supervisors' consoles and your webhooks.
Wallboards
Named layouts of tiles that rotate, with a TV mode for the floor.
Topics and contact reasons
Closed conversations tagged with a topic, a contact reason, the customer's sentiment and effort, with spikes flagged.
Revenue and cost per outcome
Deposits and conversions attributed to the last agent, channel and campaign that touched them, with the cost of the work.
Export API
Incremental CSV or JSON-lines pulls for your warehouse; a watermark means no row is repeated or skipped.
People first in performance data
Agents see their own figures and their progress towards the goals the desk sets. The leaderboard is off until the desk turns it on, and shows only agents who opted in themselves.
- Sentiment is read from the customer's messages, never inferred about employees.
- Automatic QA scores are advisory and reviewed by a person before they count.
- Retention periods per dataset, purged daily with a record of every run.
Questions about analytics
Do you analyse agents' emotions?
No. Sentiment describes the customer's side of a closed conversation. The platform does not infer emotions about employees.
Can we get the raw data into our warehouse?
Yes. The export API produces CSV or JSON lines incrementally, and carries ids rather than names, addresses or numbers.
How are figures defined?
In a versioned metric dictionary. Dashboards, reports, alerts and the API use the same definitions, so the same question gets the same number everywhere.