Private hospital
Service-desk agent for a healthcare provider.
A production deployment — running, not a pilot. The operator is anonymized at their request. Each pattern names the measure its deployment is read on; the figures themselves stay with the operator.
Last updated June 2026
The challenge
Answering which doctor is on duty now takes the receptionist away from the desk. The roster lives with the department, so she leaves the people standing in front of her and goes to ask.
What we built
A WhatsApp agent lists departments and on-duty doctors, books appointments, and states what the insurance covers and when reports can be collected. It accepts a voice note and a photograph of a prescription, because that is what patients actually send.
How it operates
Non-clinical questions are answered without the receptionist leaving her position. Appointment requests collect in one record the provider can read demand from, department by department.
The measure
The share of non-clinical questions closed without a person, and how appointment demand distributes across departments.
What happens next
The agent gives no diagnosis and no dosage. A question about a symptom leaves the agent and reaches a qualified doctor.
Deployment facts
- Channels
- WhatsApp Business API, web
- Industry
- Healthcare
- Product
- ChatSheba
- Region
- Middle East and Gulf
The operator is anonymized at their request. Each pattern names the measure its deployment is read on; the figures themselves stay with the operator.
Across every deployment
- 30+production deployments
- 35+AI agent personas
- 14+industries
Counted across every production deployment, in regulated and commercial sectors alike.