What a 1,000-call AI campaign output actually looked like

In one anonymized historical Kolz platform output, all 1,000 rows reached an ended state: 275 answered, 517 reached voicemail, 149 were not answered and 59 were cut short. Transcripts and scores were present on 846 rows.

1,000 ended rows

Dial disposition breakdown

846 rows included transcripts and scores

Voicemail517 · 51.7%

Answered275 · 27.5%

Not answered149 · 14.9%

Cut short59 · 5.9%

What this sample cannot prove

The source did not contain outcome, interest-level or success-criteria fields. This sample therefore does not show qualified leads, bookings, customers or ROI. It shows the operational output a team should inspect before claiming business results.

Campaign measurement checklist

Success event

Define one observable result, such as a requested callback or a completed qualification step.

Structured outcome

Store the result separately from dial status so an answered call is not mistaken for success.

Interest level

Use a consistent scale and keep the transcript available for review behind the score.

Opt-out status

Keep opt-outs separate and suppress them from future campaign lists.

CRM handoff

Track what happens after the call, including completed callbacks and verified pipeline changes.

How to read campaign reports honestly

Dial status explains what happened to an attempt. A transcript explains what was said. A score helps a team sort conversations. None of those fields alone proves a booking, customer or financial return. Connect a defined call outcome to the later CRM event before reporting business impact.