Montana
Call scoring
Every qualifying call is transcribed, scored and broken into fields you can filter and report on. The point is not the transcript. It is that a hundred calls become a number you can act on.
Which calls get reviewed
Not all of them, on purpose. A call is reviewed when its outcome suggests a real conversation happened, or when it lasted long enough to be one. A fifteen second wrong number is not worth transcribing and you should not be charged for it.
A call that has already been reviewed is never reviewed twice, even if the webhook fires again.
What comes out
| Field | What it holds |
|---|---|
| Summary | What happened, in a few lines. |
| Reached role | Who actually picked up: the decision maker, a gatekeeper, somebody else. |
| Objections | Each push-back, as its own entry, so they can be counted across calls. |
| Next step agreed | What was actually committed to, if anything. |
| Contact captured | Names, emails or numbers given on the call, ready to save to the account. |
| Sentiment | How the prospect sounded. |
| Topics and tags | What the conversation covered. |
| Call score | A number out of 100, with the breakdown behind it. |
How the score is built
The score is against a rubric, not a vibe. A rubric is a set of things a good call on this campaign does: opened properly, reached the right person, handled the objection, asked for the next step. Each is judged separately and the breakdown is stored, so a rep can see the two points they dropped rather than being handed a 72 with no explanation.
Rubrics are per campaign, because a call selling one thing is not a good call selling another. The default is sensible and most teams edit it once.
Using the output
Objections are the piece most teams get value from first. When the same push-back shows up in forty percent of calls, that is a script problem or a targeting problem, and it is visible in a week rather than a quarter. See Conversion rates.