Quiz & Diagnosis, playing
Salon Clair is an esthetic salon with four courses and a thousand-odd LINE friends. Pick a scenario, then take over — tap the quick replies, or type something the script never anticipated and watch what the bot does with it.
Everything here runs on scripted mock data. No message reaches LINE, and nothing you type is stored.
What each scenario is showing
The quiz as it is meant to run: four questions, a branch at the first answer, a result naming one course, and a label written against the guest. Watch the two KPI tiles move on the same turn the card appears.
She never opens the menu — she types “pores”. The bot treats that as the first answer already given and starts at question two. This is the case a web survey cannot serve, because it never knew who was typing.
“Is that course ok while I’m pregnant?” is answered from the notes on that course and nowhere else, the earlier label is corrected rather than added to, and the booking is held until a therapist has spoken to her.
The salon closed at 20:00. The gift question is answered, the certificate is issued, and the owner gets exactly one email — because notifications are set to inquiries only, and the quizzes finished that evening sent nothing.
Build this one in five minutes.
The quiz playing above is a three-question, two-branch flow. That is roughly ten minutes of typing, or one press of the generator.