CareLine WhatsApp Assistant, Healthcare clinics
Clinic WhatsApp inquiries turned into ready-to-act staff handoffs
A WhatsApp workflow for clinic inquiries that turns approved questions and intake into a useful staff handoff.
Built for: A healthcare clinic receiving patient inquiries over WhatsApp.
- intake steps before handoff
- 5
- intake steps before handoff
- clinical decisions made by AI
- 0
- clinical decisions made by AI
- safeguards: signature checks and model fallback
- 2
- safeguards: signature checks and model fallback
First agent live in one week. A short call. Your first agent, a fixed quote, and the timeline. No obligation.
Sound familiar?
- “Reception reads every WhatsApp thread to work out what the patient wants.”
- “The same service and location questions arrive all day.”
- “Contact details and appointment preferences get lost in long chats.”
How the system works.
- Step 1
Approved answers for routine questions
Replies to non-clinical questions about services, locations, availability, and preparation.
- Step 2
Structured intake
Captures contact details, the reason for the inquiry, and preferred appointment or callback times.
- Step 3
Routing with a clear reason
Sensitive, uncertain, or out-of-policy messages go to staff with a summary, the captured fields, and a handoff reason.
- Step 4
Secure WhatsApp connection
Meta WhatsApp webhooks with HMAC signature checking and replies through the Graph API.
- Step 5
Booking boundary kept explicit
A requested time stays a request until the scheduling workflow accepts it, so no slot is confirmed by mistake.
What changed.
- Every inquiry reaches staff as structured intake with a defined next action.
- Routine non-clinical questions answered without waiting for reception.
- Clinical decisions always stay with clinic staff.
- Full conversation context retained so staff pick up exactly where the chat left off.
The AI inside.
OpenAI
Writes short, channel-appropriate replies, with a fallback path if the model service is unavailable.
The full stack.
Frontend
- Next.js
- React
- TypeScript
Backend
- Webhook handlers
- Transcript handling
AI and automation
- OpenAI reply generation
Integrations
- Meta WhatsApp Graph API
- HMAC webhook signature validation
This fits you if...
- Your front desk spends hours answering the same messages.
- Patient or customer requests arrive in chat threads rather than forms.
- You need AI to handle routine questions but hand anything sensitive to a person.
Start with a free Agent Audit. First agent live in one week
The AI Operations Assistant brings this to your team.
An AI team member that handles the admin work across your email, documents, CRM, and spreadsheets, and asks before anything important goes out.
Full project write-upRead itClose
Project overview
Clinic inquiries organized for the person who handles the next step
Services: AI Receptionists ; AI Workflow Automation ; AI ERP & CRM
Delivery: WhatsApp inquiry, intake, and staff-handoff workflow. Portfolio alias.
The Problem
Clinic WhatsApp conversations mix several requests: service information, appointment preferences, intake details, and questions that require staff attention.
If those messages remain an inbox, the receptionist must read the thread, recover the contact details, identify the request, and copy the context into the next workflow. Repeated questions also compete with messages that need a person.
What We Delivered
CareLine answers approved non-clinical questions, collects contact and intake information, and prepares the next step for staff.
The workflow covers:
- Service, location, availability, and preparation questions.
- Patient contact details and the reason for inquiry.
- Preferred appointment or callback times.
- Routing of sensitive, uncertain, or out-of-policy messages.
- A conversation summary, captured fields, and a handoff reason.
A booking request and a confirmed appointment are kept distinct. The assistant should not confirm a live slot before the scheduling workflow accepts it.
Messaging Implementation
The messaging foundation uses Meta WhatsApp webhooks, incoming text handling, HMAC signature checking, and outbound replies through the Graph API.
OpenAI generates short channel-appropriate replies, with a fallback path when the model service is unavailable. Transcript entries preserve incoming and assistant messages for the application view.
The demonstration covers the conversation and handoff. Scheduling acceptance and durable CRM records are separate integrations.
Engineering Decisions
Capture the request before handing it over. The next staff member needs the reason for contact, relevant details, and the requested next action.
Keep clinical decisions with staff. The workflow does not diagnose, prescribe, or replace emergency care.
Verify channel events. Webhook authenticity checks are separate from AI reply generation.
Keep the confirmation boundary explicit. Collecting a preferred time does not mean the slot is booked.
Results
The delivered workflow turns the inquiry into structured intake and a defined next action for staff. It also retains the conversation context needed to continue the exchange.
No response-time, booking-conversion, or staffing-cost figure has been supplied. Those measurements should distinguish automated replies, staff handoffs, booking requests, and confirmed bookings.
Technology stack
Meta WhatsApp Graph API; Next.js, React, TypeScript; OpenAI reply generation; webhook signature validation and transcript handling.
Example value calculation
Hypothetical planning example, not a measured project result or a performance forecast. All volumes and timings below are assumed.
Capture contact details and the request, then review the summary and next step.
- Workload: 300 eligible non-clinical inquiries per month.
- Manual handling: 5 minutes per item.
- Assisted handling, including human review: 2 minutes per item.
- Additional exception handling: 90 minutes in the same period.
- Modeled difference: 300 x (5 - 2) - 90 = 810 minutes (13 hours 30 minutes) per month.
The example reserves another 90 minutes per month for incomplete intake and routing corrections. Clinical conversations and appointment fulfillment are excluded. This is potential staff capacity, not cash savings; setup, training, and software costs are excluded.
How to validate: Measure staff handling minutes for eligible administrative inquiries, including summary review and rework. Track handoff completeness, misrouting, booking requests, and confirmed appointments separately.
Related case studies.
Want this for your team?
Book a short audit call and show us the workflow. You leave with the first agent we would build, a fixed quote, and the timeline.