ICP Scoring Model
Scores leads on observable signals and evidenced conversion, never on rep enthusiasm or a demo request alone.
Free — no card needed
Create a free account
to open ICP Scoring Model — and the other 364 prompts across 21 categories.
Already have an account?
Running it, start to finish
- Compare won against lost characteristics before assigning any weight.
- Score fit and intent separately.
- Backtest on deals the model has not seen.
What you get back
The output this produces, every time.
- Separates fit from intent, which one combined score cannot distinguish.
- Builds negative scoring, which saves more time than positive scoring wins deals.
- Backtests against historic deals, the only real validation of a scoring model.
Getting better results
Where this usually goes wrong, and how to avoid it.
- Drop signals that appear in losses too. If a trait shows up in won and lost deals equally, it predicts nothing however sensible it sounds.
- Sample what the model rejects. There is no feedback from a lead nobody contacted, so a model's false negatives are invisible unless you look deliberately.
- Calibrate to capacity. A model producing more high scores than the team can work has ranked nothing and changed nothing.
More Sales prompts
All Sales →- Sales
Discovery Call Question Set
Builds a discovery script that finds out whether there is a real deal here, including the questions that would disqualify it early.
ClaudeChatGPT000Open → - Sales
Demo That Sells
Demos the two things that solve their stated problem, since a feature tour loses the room and teaches nothing.
ClaudeChatGPT000Open → - Sales
Cold Outreach That Earns a Reply
Writes a short first-touch built on something specific and current about the recipient, with a low-friction ask and a genuine reason to reply.
ClaudeChatGPTGemini000Open →
Written for The AI University. Every prompt in this library is original work — authored, tested and revised here, not collected from elsewhere. 365 of them, free with an account.