Script
Which version of your script the agent follows.
How Scouq tests your sales calls.
One change at a time, judged by the money each lead brings in, and only with your approval.
What Scouq can test.
Six things, one at a time.
Opening line
The first thing the agent says. It must still say it is an AI, that the call is recorded, and how to stop.
Offers
Which offer the agent presents first, and in what order the others follow.
Upsells
Which upsells the agent may offer after a yes, or none.
Payment plans
Which way to pay it offers first: in full, a deposit, or installments.
Voice
Which of your voices the agent speaks with.
The loop, step by step.
Two places where you say yes, and one safety check in between.
Step 1: Scouq proposes a test.
It picks one thing to change and writes the versions. Each version is checked first against your real offers, voices, and script. A new script cannot invent a figure or promise a result, and an opening must keep its disclosures.
Step 2: You approve the start.
Nothing runs on your leads until an owner on your team says yes. You get an email when a test is waiting.
This is the first approval.
Step 3: Each lead gets one version.
The leads are split across the versions. A lead keeps the same version for the whole test, even if they are called twice.
Step 4: Scouq measures the money.
A lead counts once they have had a real conversation. Test calls do not count. The main measure is gross profit per lead: the cash collected from that lead, times the margin on what was sold, after refunds and disputes. Installments count as they are paid.
Step 5: A safety check keeps it honest.
No winner is named until every version has enough leads, the data is complete, and one version is clearly ahead. Until then the test keeps collecting.
Step 6: You approve the winner.
Scouq asks you to make the winner your default. You can also keep things as they are.
This is the second approval.
Step 7: The winner becomes the new setup.
A winning script becomes a new approved version of your script. A winning voice becomes your default voice. Offers, upsells, and payment plans update the same way. For an opening line, a person on our team sets the new default and confirms.
The safety check against calling a winner early.
A small pile of sales can look like a pattern. The safety check makes Scouq wait for enough evidence.
30+
leads per version
Every version needs at least 30 leads that had a real conversation. A test can ask for more, never fewer.
95%
clearly ahead
Scouq re-runs the numbers many times with random re-draws of your real leads. The leader has to come out on top at least 95 times in 100.
0
winners from thin data
If margins are missing, or a version broke a limit on handoffs or refunds, there is no winner. The test keeps collecting.
Your approval, twice.
You see the numbers, then you decide.
When there is a winner, your dashboard shows both versions side by side, with leads, sales, and profit per lead. You choose one of two buttons: make the winner your default, or keep things as they are.
What a test never does.
The limits are built into the system.
- Change your live calls before you say yesA test starts only after an owner approves it.
- Name a winner earlyFewer than 30 leads on any version means no winner.
- Run two tests on the same thingOnly one test can run on the script, or the voice, or the payment plan, at a time.
- Drop the disclosuresAn opening line that does not say it is an AI, that the call is recorded, and how to stop cannot be tested.