Conversational AIConcept
Pera AI concierge
An AI assistant for a dental clinic that answers on WhatsApp and web chat, books appointments and hands over to a person when it should.
Inside the work
One conversation, two sides.
The patient's chat on WhatsApp, and the front-desk view with the CRM record and the handover.
Overview
Pera is a concept for a fictional dental clinic in Istanbul with local and international patients. Most messages ask the same few questions — prices, directions, availability — and arrive when the front desk is busy or closed. The brief: answer them well, at any hour, without ever pretending to be a dentist.
The brief
- Problem
- Patients ask the same questions at all hours; the front desk answers them one by one.
- Constraint
- It must never give medical advice, and a person must always be one step away.
- Answer
- An assistant grounded in the clinic's own content, with booking, handover and a CRM record for every conversation.
A clinic's inbox is mostly repeat questions, and a slow answer often means a patient books somewhere else. A generic chatbot makes it worse: it guesses, it loops, and it never lets you reach a person.
Pera answers only from the clinic's own content — treatments, price ranges, opening hours, aftercare notes and the booking calendar. When a question falls outside that content, or touches anything clinical, it says so plainly and passes the conversation to the front desk with the full history attached.
The same assistant runs on WhatsApp and on the clinic's website. It books into the clinic's calendar, sends confirmations and reminders, and writes every conversation to the CRM, so the team sees one patient record instead of three inboxes.
About this concept
This is a concept build for a fictional client. It shows the process and the standard we bring to client engagements.
Approach
Helpful by default, careful by design.
Grounded answers
The assistant answers from a knowledge base the clinic owns and reviews. No source, no answer — it asks a person instead.
Handover by design
Clinical questions, complaints and anything uncertain go to the front desk with a short summary, so no patient has to repeat themselves.
Connected, not bolted on
Calendar, CRM and messaging are wired in through their APIs. A booking made in chat is a booking in the clinic's own system.
What it ships with
An assistant a clinic can put its name on.
Figures refer to the concept brief.
Channels
2channels, one assistant
WhatsApp and web chat share the same knowledge, the same rules and the same CRM record.
Grounding
Answers from the clinic's own content
Every answer comes from reviewed clinic content: treatments, price ranges, hours and aftercare.
Handover
A person, one tap away
The front desk takes over with the full conversation and a summary; the patient never starts again.
Languages
3languages
Turkish, English and German, detected from the first message and kept for the whole conversation.
Stack
- WhatsApp Business Platform
- Web chat
- LLM with retrieval
- Calendar API
- CRM API
- Audit log
Try it yourself
A working build, not a picture.
Open the demo and use it the way a guest or a client would. Every tap and tab works.