AI That Answers 800 Patient Queries a Day
How a multi-city diagnostic clinic chain deployed an AI assistant that handles 80% of patient queries without human intervention
80%
Queries resolved by AI
< 30s
Avg response time
24/7
Patient support coverage
₹1.9L
Monthly cost saved
The Challenge
MediScan Diagnostics was receiving 800–1,000 patient queries every day across WhatsApp, their website, and phone calls. Their 4-person front desk team was overwhelmed — average query response time had stretched to 3–4 hours, leading to patient drop-offs and negative Google reviews. They were spending ₹2.8 lakhs per month on front-desk staff just for query handling and appointment booking, and still couldn't keep up.
Core Pain Points:
- ✗800+ daily queries — 70% of which were the same 12 questions
- ✗3–4 hour response time causing patients to book with competitors
- ✗No 24/7 availability — queries received after 8 PM had no response until next morning
- ✗Appointment booking required a phone call — no self-serve option
- ✗Report delivery notifications were manual — staff called each patient individually
""We were losing patients to competitors not because of our quality — our quality is excellent — but because no one picked up the phone fast enough."
Our Solution
Muffincode designed and deployed a custom AI-powered patient assistant — integrated into WhatsApp Business API and the clinic website — capable of answering queries, booking appointments, and sending automated report notifications.
- ✓Built a custom AI chatbot using GPT-4o with a healthcare-specific system prompt and knowledge base
- ✓Integrated with WhatsApp Business API — patients interact naturally in Hindi or English
- ✓Appointment booking flow: patient selects branch, test, date, and time slot within WhatsApp
- ✓Calendar sync with the clinic's Google Calendar — real-time slot availability
- ✓Automated appointment reminders: 24 hours before and 2 hours before via WhatsApp
- ✓Report-ready notifications: automated secure PDF download link when report is uploaded
- ✓Escalation logic: if the bot cannot answer confidently, it flags a human agent
Detailed Impact
| Metric | Before | After Muffincode |
|---|---|---|
| Query Response Time | 3–4 hours | Under 30 seconds |
| AI Query Resolution Rate | — | 80% (no human needed) |
| Daily Appointments Booked | 45 (phone only) | 120 (60% via bot) |
| Patient Drop-off Rate | ~32% | ~11% |
| Front Desk Workload | 800+ queries/day | ~160 queries/day |
| Monthly Staff Cost Saved | — | ₹1.9 Lakhs/month |
Client Response
"The AI assistant has genuinely changed how we operate. Our front desk team now focuses on in-clinic patient experience instead of answering the same WhatsApp messages 500 times a day. The ROI paid for the entire project within the first month."
— Dr. Priya Nair, Director, MediScan Diagnostics
Facing a similar challenge?
Let's discuss how we can build a technical solution to solve your specific business bottlenecks.