Reading the findings, 4 min read

What you track, and how it is decided

Six standard blocks plus the questions specific to your business, proposed from your calls, confirmed by you.

Updated

Every call is read against one set of questions, called the schema. It has two halves.

The standard blocks

  • Category and outcome: booked, missed, spam, wrong service, voicemail, follow-up, other.
  • Revenue mentioned: an amount pulled from the conversation when one was said.
  • Handling scorecard: the checks that matter for your kind of call, each scored.
  • Topics: things callers asked for that you do not offer, in their words.
  • Follow-up: whether the caller was promised one, and by when.
  • Voicemail detection.

Your own questions

Typed questions (yes/no, a choice, a number, free text) that only make sense for your business. A tutoring center asks “Was a cost actually given?” A skincare brand asks “Was an alternative to a refund offered?” A freight company asks “Was a next step booked on the call?”

How the first set is chosen

After you connect a source, three short screens ask about the business. From your answers and a sample of your first calls we write a profile and propose the questions worth asking from day one. Nothing is added until you confirm. You can edit, remove or add questions first.

VersionsA schema is immutable once active. Changing a question creates a new version, and every answer records which version produced it, so a bad batch can be found and re-run.

More in “Reading the findings”

All help articles