From a salon directory to visual, AI-assisted discovery
The move to London changed what BookBook needed to offer. As founding designer, I used research to rethink discovery, help customers evaluate recommendations and carry that confidence into booking.
AT A GLANCE
My contribution. As founding designer, I replaced heavy onboarding with visual browsing and optional AI matching, then designed salon evaluation, booking and payment flows with ML and engineering.
Product milestones. AI recommendations launched; early rollout reached 2,000+ active users and first-year downloads reached approximately 300K. These reflect the wider team’s work.
01 / THE REDESIGNED EXPERIENCE
I replaced heavy onboarding with visual discovery: customers could browse nearby salons or use a reference photo for an AI recommendation, then compare salon work, prices and services before booking.
BOOKBOOK · Visual discovery
The user chooses an image, receives a recommendation and evaluates the salon’s work before opening its profile.
Swipe to explore the full diagram
Upload a photo or start from curated looks.
Replace it if needed before matching.
The waiting state explains the work in progress.
Compare the reference and the recommendation.
NO REFERENCE READY?
Confirm the selected look to use the same matching process.
MATCHING CANNOT COMPLETE
Both fallback states offer other salons.
02 / WHY WE CHANGED
London already had established booking platforms and Instagram-led discovery. In 30–40-minute interviews, I explored how people chose salons: selection took 9–15 minutes, and 85% of participants described it as stressful.
Before: preference forms, budget sliders and style categories asked for effort before showing value. After: visual browsing came first, with image-based matching optional. I accepted less upfront preference data so customers could evaluate salons sooner.
03 / EVALUATING A SALON
ML ranked salons; I designed how customers assessed the recommendation. Their reference appears beside salon work, then a profile brings services, prices, ratings and photos into the booking decision.
I also explored review topic tags to make feedback easier to scan. The archived screen shows tags above original reviews, not the full expanded-cluster interaction.
BOOKBOOK · Find a salon
Browse the map or search directly. Both routes open the same salon profile.
Find nearby salons.
Search by place or service.
Check prices and photos, then choose a service.
Optional details help customers decide before booking.
04 / FROM DISCOVERY TO BOOKING
I carried service, specialist, date and total through checkout, updating prices as selections changed. Separate confirmations distinguish paying online from paying at the salon.
I also designed conditional prepayment around no-show history. The screens show payment branches; they do not show that eligibility rule.
BOOKBOOK · Booking & payment
Both payment options reserve the same £35 appointment. The confirmation makes any amount still due clear.
Select the specialist and appointment time.
Check the service, total and contact details.
Enter payment details before the paid confirmation.
Booking stays inactive.
Add an email and phone number.
Invalid fields show an explanation.
Keep the corrected details and choose payment.
05 / PRODUCT ANALYTICS
I defined evaluation measures for discovery time, booking completion, cancellations and no-shows. The adoption milestones below describe rollout scale; they do not establish a measured improvement in those UX outcomes.
FOUNDING DESIGNER
As founding designer in a team of nine, I owned product, interaction and brand design, partnering with ML, iOS, backend and sales. I delivered discovery and booking flows, reusable patterns and Lottie animations with iOS engineers.
Early rollout reached 2,000+ active users, launched AI recommendations and secured initial funding. First-year downloads reached approximately 300K. These are distinct milestones of the wider product and growth team.