Extraction Analytics

See per-agent KPIs, charts and field distributions built from post-call extraction data, filtered by date range, plus survey analytics.

2 min readUpdated

Extraction analytics turns the fields your extraction template fills on each call into a per-agent dashboard. It shows KPI cards (for example the share of calls with buying_intent = high), charts of how values are distributed, and a breakdown for every field. Filter by date range and share the filtered view by link.

What the dashboard shows#

Element What it is
Total extractions How many calls produced extracted data in the range
KPI cards A count or percentage of calls matching a condition, such as issue_resolved = true
Charts Pie, donut, bar or horizontal bar charts of a field's values. You can change the chart type
Field distributions How often each value appears for every field in the template
Date range Filter to any period; the range is kept in the page URL

KPI conditions#

A KPI counts calls where a field meets a condition. Supported comparisons include equals and not equals, and the result can be shown as a number or a percentage.

Example KPI Field Condition Format
Hot leads buying_intent equals high Percent
Resolved on the call issue_resolved equals true Percent
Promises to pay promise_to_pay equals true Number

Open it#

  1. Open the agent.
  2. Go to Extraction analytics.
  3. Pick a date range.
  4. Change chart types if needed; your choice is remembered.

Survey analytics#

Agents that use the survey tool also get survey analytics with the answers to each question.

Regenerate extraction#

If you change the template, old calls keep their old results. Open a call and click Regenerate to re-run extraction so it counts with the new fields.

FAQs#

Does it work without an extraction template? No. Add an extraction template first.

Can I build cross-agent custom dashboards? Not yet. Extraction analytics is per agent. Export call history for custom reporting.

Can I export the data? Yes. Export call history with the extraction columns.

Was this page helpful?