Miatz playground

The Customer Who Won't Just Tell You

The customer's real truth stays hidden until you submit.

Coming to the Build-LabBYOK AIProductpm
The Demystify signature

the persona's hidden script renders on screen only after submission, right beside the learner's own conclusion — the gap between what you heard and what was true, shown as a literal two-column diff

How it works

What goes in, what comes out

What it does

Learner interviews a live LLM-played persona carrying a hidden ground-truth problem script never shown to them. Every question is scored in real time for leading-ness, and at the end the learner writes a synthesized insight that the system diffs against the persona's hidden script.

You bring

type interview questions turn-by-turn to the persona; submit a final synthesized insight or JTBD statement

You get

a leading-question tally with a highlighted transcript, plus a side-by-side of the learner's stated insight against the persona's actual hidden script with an overlap score

You control

persona archetype picker (5 seeded personas); interview length cap (8/12/16 questions); 'coach mode' flags a leading question live as you type

Where it's used

Module 'Prompting for Product Sense' L1 lab, paired with its demo per the Prove-It Loop + free diagnostic teaser

leading vs open questionsthe Mom Testconfirmation biassynthesized insight vs stated preference

Concepts click when you open the machinery.

Three labs are already live and free — the same hands-on style this playground brings to its module.