A client's hiring volume kept climbing. Their recruiting team didn't. Every new role meant more resumes than anyone could read, and good candidates went cold while they waited.
They didn't need a smarter model. They needed a pipeline. So that's what we built.
→ Sourcing from LinkedIn, Indeed and their own candidate database in one place
→ Real-time resume scoring against each role
→ Twilio SMS and voice outreach to the top matches
→ A live ranked dashboard recruiters open every morning
Time to shortlist went from 5 days to 6 hours. 800+ candidates scored every week. Zero new recruiters hired.
Same lesson as CasePath. The AI only pays off when it's wired into the workflow people already use, and when they can see why it ranked someone first.
If one repetitive workflow is eating your team's week, that's usually where an AI agent earns its keep first.
Curious where this lands for other teams. If you could hand one workflow to an AI agent tomorrow, which one would it be? For most founders I talk to it's lead qualification, support triage or document review.
For the @Lovable booking challenge, I built AlbaSmile Booking OS.
AlbaSmile is a fictional boutique dental clinic in Prishtina, Kosovo dealing with a very real problem:
Patients want to book quickly, while clinic staff lose time checking schedules, confirming appointments, handling reschedules, chasing no-shows, and trying to refill cancellations.
The problem I solved:
Turn “Can I book?” into “You’re booked” with as little manual work from the clinic as possible.
I approached both sides of the brief.
Fix the front door
A patient can:
• choose the treatment they need
• get matched with an eligible dentist
• see real available appointment times
• book a slot directly
• receive an appointment pass
• confirm attendance
• reschedule
• cancel
• add the appointment to their calendar
No receptionist back-and-forth is required for the normal booking flow.
Fix the follow-through
On the clinic side, AlbaSmile provides:
• a live booking calendar
• inbound inquiry tracking
• appointment status management
• persisted reminder and follow-up jobs
• cancellation handling
• waitlist management
• time-limited waitlist offers
• operational visibility through one command center
The part I focused on most was cancellations.
Instead of a cancelled appointment simply becoming an empty chair, AlbaSmile can reopen that time, identify an eligible waitlist patient, create an offer, and let that patient claim the appointment.
I also turned the operational workflow into part of the visual experience.
The 3D AlbaSmile system represents what is happening behind the interface:
A booking locks into the system.
A cancellation creates a visible gap.
A waitlist patient moves through the system and fills it.
The goal was not to build another beautiful appointment form.
The goal was to build a believable booking operating system that reduces the amount of coordination a small clinic has to handle manually.
Built end-to-end in Lovable using React, TanStack Start, Supabase, Three.js / React Three Fiber, GSAP, and a persisted booking + waitlist architecture.
External messaging is simulated in the challenge build, while booking, availability, appointment management, operational state, and waitlist recovery are connected to the real application data.
AlbaSmile
A booking operating system for a boutique dental clinic in Prishtina, built to reduce the manual work between patient inquiry and confirmed appointment.