Run a Van Westendorp price-sensitivity study and a MaxDiff feature-preference study against a library of AI buyer personas, each one weighing your product as a prospective buyer and answering for itself. Minutes, not weeks. Under a dollar, not fifty thousand. Everything runs on your own machine.
Here is a finished dashboard, built from a sample run on SaaSCo, a mock B2B sales CRM we use as the worked example. This is what you get at the end: price bands, segment cuts, and ranked feature preferences. Swap in your own product and the whole thing re-runs for your market.
Open the demo dashboard →Sample numbers, for illustration. Your run replaces them with the AI personas' own responses to your product.
Not much. You do not need to be technical, and you do not need to write any code.
Claude or ChatGPT, either is fine. You will paste two prompts into it and save what it gives back. That is the only "AI" part you touch by hand.
Keep all the files together in one folder. You will save two files into it and run two short commands from it.
Want to see the dashboard before you run it for real? Preview mode fills it with sample numbers, no key, no cost. Add your key when you want the AI personas to actually respond to your product.
It comes as a ZIP folder. Unzip it and open Guide.html inside,
that is your step-by-step starting point. Runs on any Mac, Python 3 is built in.
What you need: Preview mode is free with no key. An Anthropic key (2 minutes, under a dollar per run) is only needed when you want the personas to respond to your real product.