Two working products for clinics and salons: an AI receptionist that books real appointments over live phone calls, and a voice-driven charting app that keeps pace with the probe instead of the keyboard.
Real phone calls answered naturally, live availability checked, appointments booked without collisions, every conversation transcribed.
Hands-free data capture during procedures, with clinical terminology recognition and automatic staging calculations.
Day, week and month calendar views, per-practitioner availability, server-side slot locking, and full call and chat logs.
On-device storage where patient data belongs, GDPR-friendly design, and transparent logs for everything the AI does.
Periodontal charting relied on manual data entry. It needed a dedicated assistant in the room, slowed appointments down, raised sterility concerns because of keyboard use mid-procedure, and existing commercial software was expensive. The clinician wanted charting to keep pace with the probe, not the keyboard.
We built PerioPad as a fully serverless MVP on React and Node.js, using the browser's built-in speech recognition with targeted OpenAI fallbacks for clinical terminology. The app leads the clinician step by step, recognizes pocket depths, bleeding on probing and missing teeth, calculates BSP staging and grading automatically, and renders a live interactive chart. Patient data never leaves the device, storage is local only, which makes the design GDPR friendly.
A voice-driven web app that lets a dental hygienist complete a full periodontal chart hands-free. The clinician speaks the measurements, PerioPad fills the chart in real time, calculates BSP staging and grading automatically, and exports a finished PDF report in one click.
The full periodontal chart is completed by voice, completely hands-free
BSP staging and grading are calculated automatically from the data
The dental chart updates live on screen as the clinician speaks
A finished PDF report exports in one click
Patient data stays on the device, nothing is stored on a server
Recognition reliability holds at about 95 percent on clinical terms
Clinics and salons lose bookings every day to missed calls, especially outside working hours. The client needed an AI receptionist that could handle real customer calls, prevent double bookings, and manage several doctors with separate schedules across both voice and website channels. It also had to stay fully transparent, with logs and transcripts of every conversation and a clean admin calendar behind it.
We built an operational single-clinic MVP with a clear separation between voice orchestration, AI reasoning and backend booking validation. Vapi and Twilio handle real phone calls, an OpenAI-powered brain leads the conversation, and custom server logic on PostgreSQL locks time slots so collisions are impossible. The admin dashboard shows a calendar with doctor filtering, plus full call and chat logs with transcripts. The architecture is ready to scale into a multi-tenant SaaS in the next phase.
An AI receptionist that answers real phone calls and website chats for clinics and salons. It talks to customers naturally, checks live availability, books appointments without double bookings, and keeps a full transcript of every conversation, so the front desk never misses a call again.
Customers book real appointments over a live phone call with the AI
Server-side slot locking makes double bookings impossible
A calendar view by day, week or month keeps the schedule clear
Each doctor has separate availability, managed and filtered individually
A website widget takes bookings through chat and voice as well
Every call and chat is stored with a full transcript
30 minutes. We listen, ask sharp questions, and tell you honestly what's buildable in your timeline and budget.
A clear plan, milestones, and fixed scope - so you know exactly what you're getting before a line of code is written.
Weekly demos. Production-grade code. You own everything from day one.
Gabsoft is a senior European engineering team building AI-embedded software for real estate, eCommerce and logistics. With 11+ years, 50+ projects shipped and zero failed launches, we combine senior engineering judgment with AI-accelerated workflows to ship 2-3 times faster, without cutting corners on quality.