Post-call analysis is how a messy conversation becomes a usable row in a report.
Result Types
Field type controls how users should read the value.| Type | Example result | How to review |
|---|---|---|
| TEXT | Call summary or reason for calling. | Check whether the text reflects transcript evidence. |
| SELECTOR | Resolved, Escalated, Follow-up Required. | Confirm it selected one allowed outcome. |
| BOOLEAN | customer_interested true or false. | Confirm the rule for true and false is clear. |
| NUMBER | customer age or amount. | Check formatting and units. |
Where Result Quality Comes From
Bad extraction can be caused upstream.Field description
Clear field descriptions produce better extraction.
Transcript quality
If the transcript missed the fact, extraction may miss it too.
Prompt behavior
The agent must ask for the data you want extracted.
Export mapping
Exports carry extracted values into external reporting.
Why A Field Looks Wrong
The caller never answered
The caller never answered
The extractor cannot invent a value safely.
The field description was vague
The field description was vague
Rewrite the description around evidence and allowed output.
The call used an older version
The call used an older version
Check the version attached to the call before editing the current draft.
The value is excluded from custom section
The value is excluded from custom section
Some system fields such as sentiment are treated separately from custom extracted values.
Related Reading
Post-Call Analysis Setup
Define fields.
Exporting Call Data
Use fields in CSVs.
Call Data Model
Understand stored output.