Multilingual AI assistant answered 71% of patient queries automatically.

The challenge

The hospital group's front desks were fielding hundreds of calls a day — appointment bookings, report status, visiting hours, department directions, insurance questions. Lines stayed busy, roughly a third of calls went unanswered during peak hours, and every missed call was a potential lost appointment.

Patients increasingly messaged on WhatsApp instead, where replies depended on whoever was free. Clinical staff were being pulled into answering routine questions that had nothing to do with care.

The solution

We deployed a Hindi + English AI assistant on WhatsApp and the hospital website, grounded strictly in the group's own information: departments, doctors' schedules, visiting policies, test preparation instructions, and insurance empanelment lists.

The assistant books, reschedules, and confirms appointments directly against the hospital information system through n8n, sends automated visit reminders, and shares report-ready notifications. Guardrails were non-negotiable: the assistant gives no medical advice and no diagnosis — any clinical or emergency query is immediately routed to trained staff with full conversation context.

• Hindi + English assistant on WhatsApp and web, grounded in hospital data only
• Appointment booking and rescheduling integrated with the HIS via n8n
• Automated visit reminders and report-ready notifications
• Strict clinical guardrails — medical questions always escalate to staff