customer-panellisted
Install: claude install-skill talalshaikh/customer-panel
# Customer panel
Put a pitch and its prices in front of 100 simulated customers, in four segments of 25, in the order a careful researcher would. The result shows:
- who buys, and who pays;
- what one unit is worth to them before they see your prices;
- what they compare the product to;
- their objections, in their own words;
- which offers move them;
- what would make it worth more.
## What it's for, and what it isn't
AI customers are kinder than real ones, even with the calibration rules in the brief. Use the panel to find, quickly and cheaply:
- **Objections** and the words people use for them.
- **Which segment buys, and who pays.** The best buyers are often people spending someone else's money: office managers, HR, gift buyers.
- **Reference prices.** The price people have in their heads, and what sets it.
- **What moves people.** Which offers and packages work, and which additions would raise what they'd pay.
Don't treat buy rates as a sales forecast. Don't present the panel as customer research. Prices are confirmed with real money: deposits, pre-sales, a paid pilot.
## Workflow
1. Gather the facts
2. Design the segments
3. Write the brief and `panel.json`
4. Confirm the pitch with the user
5. Run the four segments
6. Tally
7. Report
8. Next round
### 1. Gather the facts
You need:
- the product as customers will get it on the day they'd pay;
- the prices, packages and offers;
- the market (place, currency, who the customer is);
- how and where it's delivered;