4.6-rated AI agents across a private hospital network in Namibia

Supporting patients through surgery – before and after

Patients preparing for a procedure had one option before the deployment: call reception during opening hours. Calls outside those hours went unanswered. Post-operatively, follow-up was ad hoc – some patients received wound care guidance and warning signs information; others did not.

Ryno and Wally now deliver pre-operative instructions at any time of day: fasting requirements, arrival guidance, what to bring, and procedural information prepared by the hospitals' clinical nursing teams. Every patient receives the same information, consistently, regardless of shift or staffing. Post-operative support works the same way – wound care guidance, warning signs, and follow-up reminders, available 24 hours a day.

Pre-admission intake is handled through the same channel. The agents guide patients through a conversational form that collects identity, admission details, medical aid information, and next-of-kin before the patient arrives. Reception receives a pre-populated ticket; patients arrive partially processed. Front-desk handling time is reduced, and queues are shorter.

99.8% of all patient conversations were resolved by the AI agents without staff intervention – against a target of 95%. The AI first response time was 4 seconds. In the first month of operation alone, automated handling displaced an estimated 50 hours of reception and administrative staff time – an early figure that reflects initial adoption levels and will grow as patient awareness of the channel increases.

Specialists, billing, and general information – on a single channel

Finding the right specialist previously required a phone call, a wait, and a staff member who could direct the patient correctly. Ryno and Wally give patients instant access to the full specialist directory across both hospitals – every consultant, searchable by name or field of specialisation, with contact details and visiting schedules returned directly in the conversation, at any hour.

Specialist lookups accounted for nearly 1 in 10 of all conversations handled in the first month – a consistent, high-frequency demand that was previously absorbed entirely by reception staff on the phone.

Billing and medical aid queries, visiting hours, parking, services, and general hospital information are handled the same way – collectively representing nearly a quarter of all conversations, fully automated.

Patients rated their experience 4.6 out of 5 – measured from real ratings within the first month, with no patient-facing marketing campaign active at one of the two hospitals for the full period.

An unexpected discovery: the recruitment use case

The agent workflows were designed around patient experience. What emerged in practice was broader.

42% of all conversations in the first month were about recruitment and job applications – the single largest category by volume, and one that was not a primary design target. In the hospitals' prior model, these enquiries were handled manually by administrative or HR staff: inconsistently, without a record, and during business hours only.

Once the pattern was identified, the recruitment workflows were extended within a single day. Ryno and Wally were updated to guide applicants through the submission process, surface which roles were actively open, and confirm which positions had already been filled – removing the need for any manual HR involvement in first-contact recruitment enquiries.

The agents now handle every recruitment conversation automatically. In a hospital group expanding across multiple facilities, the volume of HR-related enquiries will grow proportionally. The deployment already handles it.

Patient experience as a revenue lever

In private healthcare, patients choose their provider. In a concentrated market – where the same patient base selects from a small number of private hospitals – patient experience shapes that choice as much as clinical reputation.

A 4.6/5 satisfaction score, achieved without a patient-facing campaign in the first month, is evidence that patients engaging with this channel are having a positive experience. A 4-second response to a question at 11pm about post-operative wound care is an experience the hospital's competitors cannot offer. Every interaction that resolves a patient's question quickly and correctly builds the perception that informs where that patient goes next – and who they refer.

In a market this size, those marginal advantages accumulate. The hospitals that are reachable, responsive, and consistent in patient communication are the ones that hold that preference over time.

One platform, two hospitals – and an expanding network

Ryno and Wally each operate as dedicated agents for their hospital, tailored to their services and patient community. Both run within the same Proto enterprise deployment – and that shared architecture is what makes agent-to-agent handover possible across the network.

When a patient's query falls under another facility's scope, the agent transfers the conversation directly within the same WhatsApp or webchat thread, with full conversation history intact. The patient does not start again.

As the network expands to additional facilities in the group, each new agent joins the same platform. What that shared infrastructure enables across the network:

• Agent-to-agent handover – patients referred between facilities without restarting the conversation
• Specialist referrals across sites – directing patients to the right hospital for a specific procedure or consultant
• Rehabilitation and radiology referrals – routing post-surgical patients to partner facilities for recovery or imaging
• Centralised KPI tracking – automation rate, satisfaction scores, and interaction volume monitored across all agents from a single dashboard
• On-premise data hosting – patient data stored within the hospital group's own infrastructure, with AI compute in the cloud