Case study 05AI patient query desk for a medical clinic
Swastik Clinic Assistant
An AI assistant that answers patient questions instantly, so clinic staff don't have to.
A chat assistant that answers a clinic's most repeated questions around the clock, only from the clinic's own verified information, and hands everything else straight to the front desk.
View the live assistant (opens in a new tab)- Built for
- Swastik Medical Centre, Vasundhara, Ghaziabad
- Built by
- CODA
- Duration
- 1 week
- Platform
- Web, mobile-first
All project details
- Built for
- Swastik Medical Centre, Vasundhara, Ghaziabad
- Built by
- CODA
- Industry
- Healthcare · OPD · patient flow automation
- Duration
- 1 week
- Platform
- Web, mobile-first
01Overview
The same four questions, all day long.
A large share of a clinic's front-desk time goes to the same few questions. They arrive by phone and WhatsApp, often outside working hours, and they pull staff away from the patients who are physically in the clinic.
- Which doctors consult here?
- What are the OPD timings?
- What's the fee?
- How do I book?
By phone and WhatsApp, often outside working hours.
02The problem
Busy staff, waiting patients, mixed-up answers.
- 01
Repetitive workload
The same questions, answered again and again by the people who should be looking after patients.
- 02
Delayed answers
Messages sent after hours or during a busy clinic waited for someone to be free.
- 03
Inconsistent information
The answer depended on who happened to pick up the phone.
03Our approach · four steps
Design. Build. Secure. Host.
Built around what patients actually ask.
We mapped the most common patient queries and shaped a mobile-first conversational experience around them. Starter prompts cover the four things patients ask most, and WhatsApp and call buttons are always one tap away for anything the bot shouldn't handle.
What we delivered
- Query taxonomy
- Conversation UX
- Starter prompt flows
- Brand and interface design
04The conversation
Ask in plain words. Get a plain answer.
Four starter prompts cover what patients ask most. Here is a real exchange from the assistant, replayed. Watch what it does when the clinic's own information disagrees with itself.
- 1Doctors and specialitiesWho consults at the clinic
- 2OPD timingsWhen each doctor is available
- 3Consultation feesCharges for each consultation
- 4Book an appointmentHow to schedule a visit
According to the information provided, the timings of the doctors in the clinic are as follows:
Dr. Rahul Gupta
- Monday to Saturday: 10:00 AM – 1:00 PM
- Monday to Saturday: 6:00 PM – 8:00 PM
- Sunday: 10:00 AM – 12:00 PM
- Consultation fee: ₹700
Dr. Priya Gupta
- Monday to Saturday: 10:00 AM – 12:30 PM
- Consultation fee: ₹1,100 (website fee) or ₹1,200 (online booking catalog)
Please note that there is a conflict in the consultation fee for Dr. Priya Gupta, and it's recommended to confirm the current fee during booking.
Two sources gave two different fees for the same doctor. The assistant does not pick one. It shows both, flags the conflict, and tells the patient to confirm when booking.
05 · Built to be safe
In healthcare, a wrong answer has a cost.
It does not guess
If the answer is not in the clinic's information, it says so and points to the front desk.
It flags conflicts
When two sources disagree, it shows both and asks the patient to confirm, instead of quietly choosing one.
It points to emergency care
Every screen carries a reminder: AI responses may be inaccurate, and in an emergency call 112.
It keeps keys private
Credentials stay on the server, never in the page.
It stores no health data
There are no patient accounts and nothing about a patient is kept.
06Outcome
A first-line responder that never clocks out.
Swastik Medical Centre now has a 24/7 first-line responder for its most repetitive patient queries. Patients get instant, consistent answers about doctors, timings and fees at any hour. Anything beyond the bot's scope goes straight to the clinic through WhatsApp or a call, which frees management staff to focus on patients in front of them. Updating the clinic's information is a single-file edit, so the assistant stays current without developer involvement.
Always on
Answers at any hour, including when the clinic is closed.
Consistent
Every patient gets the same, verified answer.
Staff freed up
Management focus on patients in front of them.
Easy to keep current
Updating the clinic's information is a single-file edit.
- Build
- Next.jsResponsive, mobile-first layout
- AI
- Hosted LLM inference via APIKnowledge-grounded answersStreaming responses
- Hosting
- Vercel
- Focus areas
- Healthcare AIPatient engagementClinic automationConversational UIWhatsApp handoff
© 2026 CODA · Citizen Of Digital Age. The conversation above is replayed from a real exchange with the assistant.