Stress-test your idea before you build a single screen.
Paste an idea, convene a panel of evidence-grounded AI users, and get sharp objections, confusion points, and a build / refine / rethink verdict in under a minute. It red-teams your assumptions, it does not predict success.
A spread of adoption appetites
5 AI users assemble and react in seconds
The panel
Marisol Reyes
Owner-operator, hands always busy
Runs a three-chair salon in San Diego. Cuts hair eight hours a day and answers the desk phone between clients.
Wants: Stop losing bookings to voicemail and stop touching a greasy phone screen mid-cut.
Booking without putting down the scissors is the dream. If a client calls and I can just say 'book Tuesday at 3', that saves me ten interruptions a day.
I do not get whether it answers calls for me or whether I am the one talking to it. Those are very different products.
If it mishears a name or a time in a loud salon, I have double-booked a real human. One bad slot and I stop trusting it.
Save me from the phone, but do not double-book my chair.
Darnell Carter
Skeptical barbershop veteran
Owns a busy barbershop with four barbers. Tried two booking apps before and went back to a paper book.
Wants: Wants fewer no-shows, but refuses to babysit yet another piece of software.
Hands-free is genuinely new. Every other app made me stop and tap. Talking to it while I sweep up is the first thing that feels built for my floor.
Where does my existing calendar go? I have two years of regulars in another app and nobody explains the switch.
Voice in a room with clippers buzzing and music going? I will believe the accuracy when I see it survive a Saturday rush, not a quiet demo.
Show me a Saturday, not a quiet office demo.
Priya Nair
Tech-forward early adopter
Solo stylist who runs her whole business from her phone and already uses voice assistants daily.
Wants: Loves trying tools that make her look modern and save admin time at night.
This fits how I already work. I talk to my phone constantly. Booking by voice between clients feels obvious in the best way.
Can my clients book by voice too, or is it only for me? The landing idea reads both ways.
I would not pay more than my current calendar costs unless it also handles reminders and deposits. Voice alone is a feature, not a product.
Voice alone is a feature. What is the rest of the product?
Helen Whitcombe
Cautious late majority
Manages a long-standing salon with an older clientele. Not comfortable with anything that feels experimental in front of customers.
Wants: Keep the front desk calm and predictable. Avoid anything that could embarrass her with a regular on the line.
If it quietly takes a booking while I finish a blow-dry, fine. I like the idea of not making clients wait on hold.
I would worry I have to learn special commands. If it does not understand normal speech, it is just a fancier way to make mistakes.
My regulars are in their sixties. If the app talks back to them in a robotic voice, they will hang up and call me annoyed.
My regulars will hang up on a robot. Then they call me.
Theo Lindqvist
Operations-minded multi-location owner
Owns three salon locations and thinks in spreadsheets, staff utilization, and conversion rates.
Wants: Squeeze more booked hours out of each chair without hiring a receptionist per site.
If this captures the calls that hit voicemail after hours, that is real recovered revenue I can measure. That is the pitch I care about.
How does it route across three locations? A voice booking is useless if it lands a client at the wrong shop.
I need reporting. If I cannot see how many bookings voice captured versus lost, I cannot justify rolling it to three sites.
Recovered after-hours bookings? Now show me the report.
The reckoning
Panel verdict
Refine
Real pain, fuzzy product boundary.
The panel agrees the hands-free moment is genuinely novel and the after-hours booking loss is a wallet-level pain. The split is about what the product actually is: owner-side dictation or a customer-facing phone agent. Accuracy in a noisy room and migration from existing calendars are the trust gates that decide whether anyone pays.
Panel willingness
54 / 100
The three riskiest assumptions
- 1
Voice recognition stays accurate in a loud salon.
Three of five personas raised noise or mishearing as the thing that would make them quit. A single double-booked chair erodes trust faster than the feature builds it.
- 2
Owners will move their existing calendar and clients over.
Veterans with years of regulars in another tool see migration, not setup, as the real cost. Nobody in the panel believed switching was free.
- 3
Voice booking alone is worth paying for.
The early adopter and the operator both reframed voice as one feature inside a larger booking product (reminders, deposits, reporting), not a standalone purchase.
Run this next, with real people
Run a noisy-room accuracy test with five real owners
Sit in five working salons during a busy hour and have owners dictate ten real bookings each. Log the misheard rate and watch their reaction to the first mistake. If accuracy holds above roughly 95 percent in noise and they shrug off one error, the core bet survives. If one mistake makes them reach for paper, refine the trust model before building anything else.
Synthetic personas red-team your assumptions. They supplement real user research, they do not replace it.
Why a red-team, not a hype machine
Synthetic users are great at surfacing objections and comprehension gaps, and unreliable as success predictors. We built around what they are actually good for.
Distinct voices, real objections
The prompts force each persona into a different stance and require at least one hard dealbreaker. No five-way chorus of yes.
Assumptions, not predictions
You leave with the three riskiest assumptions your idea depends on and one concrete test to run with real humans next.
Honest by design
Every persona claim is tagged with its reasoning basis so you can see when it is a stated need versus a demographic guess.