WhatsApp as a hospital’s front door
WhatsApp for hospitals works when it is run as a designed front door rather than a side channel on staff phones. That means an official business account, a clear number strategy across units, a short menu for the common intents, fast handoff to trained people, strict rules on reports and clinical questions, CRM integration so every conversation belongs to a patient record, and metrics tied to appointments.
Walk through the OPD of almost any Indian hospital and you will find WhatsApp already doing front desk work. The coordinator sends a report to a family on her own phone. A doctor’s assistant confirms the evening slot from a personal number. The billing desk shares an estimate as a photo of a printout. None of it is designed, none of it is recorded, and all of it happens because patients prefer it to a call.
That preference is the point. For a large part of India, WhatsApp is the default way to talk to any business, and a hospital that only offers a phone line and a web form is asking patients to change habits to reach it. The question for a growth or digital leader is not whether to use WhatsApp for hospitals. Patients have settled that. The question is whether the hospital runs it as a front door, with the same care it gives the reception area, or keeps letting it happen in pockets.
This article covers what that front door looks like once it is designed: the account and number decisions, the conversation design, where people come in, the rules that keep it safe, and how to know it is working.
Why the unofficial version is a problem
The informal channel feels efficient. Staff are responsive, patients are happy, and nobody had to buy anything. The costs show up elsewhere.
Patient data sits on personal devices. When a coordinator leaves, the conversations, reports and phone numbers leave with her. There is no record the hospital can review when a complaint arrives, and no way to show what was said about a bill or a procedure date. Under the DPDP regime, the hospital is responsible for personal data processed on its behalf, and a report sent from a personal phone is very hard to defend.
There is also a commercial cost. Enquiries that arrive on individual phones never reach the CRM, so they are never counted, never followed up systematically and never attributed to the campaign that produced them. A unit can be receiving good demand through WhatsApp and look weak on every dashboard. I would rather have that demand visible, even if the first view is messy.
Setting up WhatsApp for hospitals properly
The foundation is the official WhatsApp Business Platform, accessed through an approved provider, with a verified business profile in the hospital’s name. The small business app used on a single phone is not built for a multi-desk operation. The platform version gives you a shared inbox for many agents, message templates approved in advance, automation, and an integration route into your CRM and appointment system.
Two rules shape everything that follows. First, the hospital can reply freely only within a service window that opens when a patient messages. Outside that window, the hospital can only send pre-approved templates. Second, business-initiated messages need the patient’s opt-in, and templates are categorised (utility messages such as confirmations and reminders are treated differently from marketing). Design the journeys around those rules from day one, rather than discovering them when a campaign is blocked.
Think about entry points at the same time. Patients will arrive from a click-to-chat button on the website, the doctor page, the Google Business Profile, a QR code at reception or in the OPD waiting area, printed prescriptions and campaign ads. Each entry point can carry a pre-filled first message or a tag, so the conversation opens already knowing where the patient came from and roughly what they want. That small detail saves an exchange of messages and makes attribution possible later.
One number or many
Multi-unit groups have to choose between a single group number and a number per unit. A single number is easier to publicise, builds one relationship with the patient and keeps quality ratings in one place. Per-unit numbers feel local and let each unit own its queue, but they fragment the patient’s history and multiply the governance work.
My preference is one group number with routing: the first question in the conversation establishes which unit or city the patient is asking about, and the conversation lands in that unit’s queue. Local teams keep ownership, the patient keeps one thread with the hospital, and the group sees the whole picture. Print that number everywhere the old personal numbers used to appear, and ask doctors and coordinators to redirect patients to it.
Designing the conversation
Most hospital WhatsApp journeys fail in the first few messages. A patient types a question and gets a wall of text with a numbered menu of a dozen options. Or they get a chatbot that answers everything with “please call our helpline”.
The better pattern is a short button menu for the intents that make up most of the traffic, and a clear route to a person for everything else. For a typical unit, the menu is small.
- Book or change an appointment
- Get my reports
- Bills, estimates and insurance
- Directions and timings
- Talk to someone
Booking should open with a specialty or doctor choice, show available slots where the scheduling system can expose them, and confirm with a template the patient can find later. Directions and timings should answer instantly with a map link. Bills and insurance nearly always need a person, so route them quickly rather than pretending a bot can explain a TPA query.
Free text deserves care. Patients will type “my mother has fever since three days, which doctor” in a mix of Hindi and English. An AI layer can classify that intent and route it well, but it should not answer the clinical part. I have written about where AI belongs in the patient journey, and on WhatsApp the line is the same: understand the request, route it, book it, never advise.
Language, tone and the written voice
A hospital’s phone scripts rarely translate well to chat. On the phone, a warm voice covers for a clumsy script. In writing, every word sits on the screen and gets forwarded to the family group. Agents need a short written style guide: greet by name, answer the question first, keep messages short, confirm what will happen next and when, and never paste a paragraph of policy.
Language matters even more. Patients will write in Hindi, Tamil, Bengali or a mix typed in Roman script. The menu and templates should be available in the main languages of each unit’s catchment, chosen by the patient at the start, and the routing should send regional language conversations to agents who can read and reply in them. A template approved only in English will be read, but it will not feel like the hospital is speaking to that family.
Keep the tone consistent with the brand, not with the individual agent. Emojis, slang and over-familiar sign-offs look friendly in one conversation and careless in a complaint review. Decide the house style once, show examples, and let supervisors review a sample of conversations every week the way they review call recordings.
Where people take over
The handoff to a human is where the channel earns or loses trust. A patient who chooses “talk to someone” and waits in silence will call instead, and will remember the silence.
That means staffing the inbox as a real queue, inside the contact centre or a central patient desk, with named agents, working hours stated in the automated reply, and a response time target the team actually tracks. It means agents see the patient’s history (previous visits, open appointments, earlier conversations) before they type, which requires the CRM integration. And it means a clear rule for what the bot says outside staffed hours: when a person will reply, and what to do if the matter is urgent.
The judgement about when a machine should step back is the same one I described for voice in when automation should not answer the phone. Distress, complaints, billing disputes and anything clinical should reach a person early. A bot that measures success by how many conversations it kept away from staff will look good on a slide and poorly with patients, which is the trap covered in containment versus resolution.
Rules that keep the channel safe
A few rules should be written down, approved by the medical director and legal, and built into the tooling rather than left to judgement.
Emergencies. Any message that looks urgent gets an immediate automated response with the emergency number and ambulance route, and is flagged to a supervisor. The bot must never make a patient wait in a queue for that.
Clinical questions. Agents do not give medical advice on WhatsApp. They book a consultation, arrange a call-back from a clinical team where the hospital offers one, or direct to emergency care. Doctors who want to message their own patients should do it from an approved hospital account, with the conversation recorded.
Reports and documents. Sending lab or imaging reports is one of the most valued uses, and one of the riskiest. Verify identity before release, prefer a secure link with a one-time code to an attachment, and never send reports to a number that is not registered to that patient without a verification step.
Marketing. Promotional messages go only to patients who opted in for them, at a sensible frequency, with an easy opt-out. Blasting health check offers to everyone who ever enquired is the fastest way to get the number’s quality rating downgraded and messaging limits reduced.
Connecting it to the CRM and the appointment system
WhatsApp without integration is a nicer phone line. With integration, it becomes part of how the hospital knows its patients. Every conversation should be linked to a patient or lead record, tagged with intent and unit, and carry its source (a campaign link, a QR code at a unit, the doctor page, a Google profile).
That is what lets the channel do the useful work around a visit: appointment confirmation, a reminder the day before, directions on the morning, a feedback request after, and a follow-up prompt when the doctor has asked the patient to return. Each of those is a utility template triggered by the appointment system, not a campaign someone schedules by hand. For groups deciding whether they are ready for this, the hospital CRM readiness checklist is a sensible first pass, and the case for the CRM itself is in what a hospital CRM is actually for.
How to measure a front door
The early reports will be full of volume: conversations started, messages exchanged, menu clicks. Those tell you about adoption, not value. The measures I would put in front of a unit head are simpler.
How many conversations with booking intent became appointments, and then completed visits. How long a patient who asked for a person waited for the first human reply, in staffed hours and outside them. How many conversations ended unresolved or moved to a phone call. How many patients opted out of messages or blocked the number, and what they had received just before. And, over time, how much of the informal traffic on personal phones has moved to the official number.
Read them by unit. One unit with slow human response will drag down the group number’s reputation for everyone, and the unit head should see that in their own review.
What I would switch on first
Resist the temptation to launch every use case at once. A front door that does three things well is better than one that does ten things badly.
Start with the official account, the single number with unit routing, and a staffed inbox during OPD hours. Add appointment confirmations and reminders as utility templates, because they are low risk, useful to every patient and quick to show value. Put the new number on signage, doctor pages, prescriptions and the Google profile, and ask staff to move patients off personal phones as the first operational rule.
Once the handoff works and response times are steady, add report delivery with proper verification, then booking inside the chat. Marketing messages come last, and only to patients who opted in. At each step, check the numbers that matter: appointments and completed visits, human response time, opt-outs. Run it this way and WhatsApp stops being something the hospital tolerates and becomes the front door patients already wanted.
Questions people ask
WhatsApp for hospitals means using the official WhatsApp Business Platform as a designed patient channel: booking, confirmations, reminders, reports, directions and billing queries, handled through a shared team inbox, connected to the CRM and appointment system, and governed by written rules. It replaces the informal version, where staff use personal phones, with a front door the hospital can staff, record and measure.
Because the hospital cannot see, record or protect what happens there. Patient data sits on personal devices, conversations leave with staff, complaints cannot be reviewed and enquiries never reach the CRM. Patients may be happy on the day, but the hospital carries the privacy risk and loses the demand data. The official channel keeps the responsiveness patients like while fixing those gaps.
It can, but I prefer one group number with routing by unit or city at the start of the conversation. The patient keeps one thread with the hospital, the quality rating sits in one place and the group sees all demand. Each unit still owns its own queue and response times, so local accountability is not lost by centralising the number.
No. A bot, or an AI layer, can recognise that a message is about a symptom and route it to booking, a clinical call-back or emergency guidance. It should not give medical advice. Agents follow the same rule. The medical director should approve this boundary in writing, and the automated responses should be reviewed regularly to check they stay on the right side of it.
It can be, with controls. Verify the patient’s identity before release, prefer a secure link protected by a one-time code rather than an attachment, and never send reports to an unregistered number without verification. Record every release against the patient record. Report delivery is one of the most valued uses of the channel, so it is worth doing carefully rather than avoiding.
The WhatsApp inbox has to be staffed and managed as a real queue, with named agents, stated hours, response time targets and supervisors watching the backlog. Agents need training in written tone, which differs from phone scripts, and access to patient history in the same screen. Many groups run it from the contact centre because the skills, rosters and quality processes already exist there.
Business messages need a lawful basis, which for marketing usually means clear consent recorded at the point of opt-in, with an easy way to withdraw. The hospital must know what data flows through its provider, where it is stored and how long it is kept. Written agreements with the platform provider, and a review by the privacy lead, should come before launch, not after.
Look at integration with your CRM and appointment system, the quality of the shared inbox for many agents, routing by unit, template management, data storage location and retention controls, audit logs, and support responsiveness. Ask how conversations are exported if you ever switch provider. Avoid choosing on message pricing alone, since integration and inbox quality drive most of the real value.
Yes, but only to patients who opted in to promotional messages, at a sensible frequency, with an easy opt-out. Health check or camp announcements sent to everyone who ever enquired will generate blocks and complaints, which can lower the number’s quality rating and restrict messaging for the whole hospital. Keep utility messages such as reminders separate and protect them.
A basic launch with the official account, a number with unit routing, a staffed inbox and appointment reminders can be done in a matter of weeks if the CRM and appointment system are accessible. Report delivery and in-chat booking take longer because they need deeper integration and sign-off. Plan in phases and let each phase prove itself before adding the next.
Appointments and completed visits that started on WhatsApp, the time patients wait for a human reply, unresolved conversations and opt-outs, all by unit. Also ask how much informal traffic on personal phones has moved to the official number. Conversation volume alone is not a success measure. It shows adoption, not whether the channel is converting demand or serving patients well.
Doctors who want to stay in touch with their patients should do it through an approved hospital account or through the central team, so conversations are recorded and protected. Many will prefer to keep personal numbers, and change takes time. Start by redirecting new patients to the official number and giving doctors a quick way to ask the team to send confirmations or follow-up reminders.
Launching a bot that tries to keep patients away from people. Long menus, circular answers and endless requests to call the helpline frustrate patients who chose WhatsApp to avoid calling. The channel works when automation handles the simple, repeatable requests quickly and hands everything else to a trained person without delay, with the patient’s history already visible.
