Session 3 · team build

Make the AI score calls the way you sell.

A stock prompt gives stock output. Claude interviews your team about how you actually sell, rewrites the scorecard for you, then proves the new one is sharper by running both side by side. Generic in, generic out.
  1. Copy the prompt below and paste it into a new Claude chat.
  2. Work through it as a group. One person types, everyone talks.
  3. Push back on it. If the output is generic, tell it so and make it try again. The argument is the learning.
  4. Save what you make. It is your team's, from today.
Paste into Claude
Help my team turn a generic call scorecard into one that fits how we actually sell.

Interview me one question at a time, up to six. Wait for my answer each time.
1. Who do we sell to, and what makes an account a good fit for us?
2. What are the stages of our sales process, in our words rather than textbook ones?
3. What does our team consistently get wrong in discovery?
4. What has to be true before we would call a deal properly qualified?
5. Where does the scored output need to end up: CRM, a deal review, a manager one to one?
6. What does a generic scorecard get wrong for us?

Then rewrite the scorecard prompt for my team. Tell me what you changed and why, line by line, so I can defend it to my manager.

Then prove it. Write a short sample call, run both the generic version and your new version against it, and put the two outputs side by side so we can see exactly where the new one is sharper. If it is not sharper, say so and fix it.

Finish with the final prompt in one clean block I can copy, and one sentence on how we should keep tuning it as we learn.
What your team brings back

Your team's own version of the scorecard prompt, tuned to your ICP and your process, with proof it beats the generic one.

Winning by Design · AI for GTMSession 3 · Run the Call, Prove the Impact