AI and automation for hospital and clinic operations in India

The problem

The front desk phones patients the evening before. Many calls go unanswered, and no-shows leave consultants idle while walk-ins wait.

What we build

A booking in the HMIS triggers a WhatsApp confirmation, a reminder the day before and one two hours before, with Confirm and Reschedule buttons. The front desk sees responses in one list.

Expected outcome

Typically, confirmation rates rise within the first month and the front desk stops the evening call round; the gain shows up as fuller OPD slots and shorter waits.

How it works

1. Discovery: we map how appointments enter the HMIS today (phone, walk-in, website, aggregator apps) and which consultants run on slots versus tokens.
2. When the receptionist books a slot in the HMIS, n8n picks it up through the API, a database view or a scheduled export and sends a WhatsApp confirmation with date, consultant and directions.
3. Reminders go out 24 hours and 2 hours before. Patients tap Confirm or Reschedule; reschedule requests land in the front-desk queue.
4. Each morning the front-desk lead gets the confirmed, unconfirmed and cancelled lists on WhatsApp so open slots can be offered to the waiting list.
5. Patients who missed receive a polite rebooking message the next day; the OPD coordinator sees who has not rebooked after 7 days.
6. A monthly report per consultant shows booked, confirmed, arrived and no-show counts, with messages sent and their cost.

Systems: HMIS · WhatsApp Business API · n8n · Google Sheets

The problem

Every patient fills the same form at the counter. Queues build at 9 am, and staff retype details into the HMIS while the phone rings.

What we build

A QR code at the entrance or in the reminder message lets patients pre-register on their phone. They get a token on WhatsApp, and a display shows who is next. Where the hospital is on ABDM, the same flow can use Scan & Share.

Expected outcome

Typically, counter time per patient falls to under a minute and the administrator can see the morning queue instead of guessing it.

How it works

1. The receptionist and OPD coordinator walk us through the morning rush; we count the steps from the door to the consultant's room.
2. A short mobile form (name, phone, consultant, ABHA number if the patient has one) is linked from a QR code at the entrance and from the reminder message.
3. The form creates or matches the patient record in the HMIS; the patient receives a token number on WhatsApp.
4. A screen at the counter and a TV display show the live queue per consultant; the nurse station calls the next token.
5. Walk-ins without a phone are added by the receptionist in the same screen, so there is one queue.
6. Daily statistics: average wait per consultant, peak hour, and tokens that left without being seen.

Systems: HMIS · WhatsApp Business API · QR codes · ABDM Scan & Share · n8n

The problem

Patients phone or come back to ask whether their report is ready. Staff print, search and hand over, and reports go missing.

What we build

When the lab marks a report final in the LIS or HMIS, the patient receives a PDF on WhatsApp with recorded consent, and the referring consultant's desk is notified. No interpretation is added.

Expected outcome

Typically, report-status calls to the lab counter drop sharply and reports reach patients the same day they are signed off.

How it works

1. The lab in-charge and the IT person map how a report becomes final today and where the PDF is stored.
2. Consent to receive reports on WhatsApp is recorded at registration and the mobile number is verified by OTP.
3. n8n watches the LIS or HMIS for reports marked final and sends the PDF, or a time-limited link, with an approved template message.
4. Reports the lab marks as urgent are routed to the consultant's assistant or the ward nurse station as a notification so follow-up is not missed. The message carries no clinical comment.
5. Delivery status (sent, delivered, read) is logged against the patient record; undelivered reports appear on the lab desk list.
6. Patients can reply with their token number to get a report resent; the bot handles it without staff.

Systems: LIS · HMIS · WhatsApp Business API · n8n · secure file storage

The problem

Pre-authorisation forms, bills, discharge summaries and ID copies are assembled by hand for every insured patient. TPA query emails get lost in a shared inbox, and money stays stuck.

What we build

A claim tracker that opens a file per insured admission, builds the document checklist for that TPA, pulls bills and summaries from the HMIS, logs every query and shows days outstanding per claim.

Expected outcome

Typically, the insurance desk stops rebuilding files from scratch and the owner can see outstanding claim value by TPA on any day, not at month end.

How it works

1. The insurance desk and the accounts lead list the TPAs, their document requirements and where delays happen today.
2. When an insured patient is admitted, a claim file opens with that TPA's checklist; missing documents show in red.
3. Bills, the discharge summary as issued by the treating team, and ID proof are pulled from the HMIS and attached; the desk uploads the pre-auth response.
4. TPA query emails are read and attached to the right claim; the desk gets a WhatsApp nudge with a due date.
5. The administrator's dashboard shows claims by stage, amount outstanding, days pending and TPA-wise rejections.
6. A monthly export goes to Tally Prime or Zoho Books for receivables and write-offs.

Systems: HMIS · TPA portals · Gmail or Outlook · Google Drive · Google Sheets · WhatsApp Business API · n8n · Tally Prime

The problem

Near-expiry stock is found at the annual count. Wards indent on paper, purchase approvals wait for the administrator's signature, and some items run out mid-week.

What we build

Expiry and reorder alerts from the pharmacy module, ward indents from a simple form or WhatsApp, and a one-tap purchase approval for the administrator that posts to Tally Prime.

Expected outcome

Typically, expiry write-offs and emergency local purchases fall within a quarter, and ward indents stop being a paper trail.

How it works

1. The chief pharmacist and the purchase officer set minimum stock, supplier lead times and expiry thresholds per item group.
2. Each night, n8n reads the pharmacy module (inside the HMIS or standalone) and lists below-minimum items and batches expiring in 60 and 30 days.
3. Ward nurses send indents from a form or WhatsApp; the pharmacy sees them in one queue and issues stock with barcode scanning.
4. Draft purchase orders per supplier go to the administrator on WhatsApp; approval posts the PO to Tally Prime and emails the supplier.
5. Goods receipt is scanned in with batch and expiry; short supplies are flagged to the purchase officer.
6. A monthly report shows consumption by ward, expired value and the top stock-outs.

Systems: HMIS pharmacy module · Tally Prime · barcode scanners · WhatsApp Business API · n8n

The problem

Discharge waits for the bill. Charges sit on paper sheets that do not match the HMIS, and consultant share is worked out in Excel at month end, with disputes.

What we build

A daily interim-bill check, a discharge checklist that runs across departments, and consultant share statements calculated from coded rules and posted to accounts.

Expected outcome

Typically, discharge time shortens and month-end consultant payout disputes fall, because every figure comes from one source.

How it works

1. The billing head, the nursing superintendent and accounts map the discharge path and where charges are missed: OT consumables, procedures, pharmacy returns.
2. Each morning the system compares HMIS charges with ward and OT records; gaps go to the billing desk before the patient is due to leave.
3. When the treating team marks a patient fit for discharge, a checklist runs: pharmacy returns, pending reports, bill ready, TPA pre-auth status.
4. The patient's attendant receives the estimated bill and payment options on WhatsApp; cash, UPI and card receipts reconcile to Tally Prime.
5. Consultant share rules (per procedure, per consult, per package) are coded once; monthly statements go to each consultant with line-level detail.
6. The administrator sees average discharge turnaround and billing leakage by department.

Systems: HMIS · Tally Prime · WhatsApp Business API · Google Sheets · n8n · UPI and card payment gateway

The problem

OPD count, bed occupancy, the OT list, collections and outstanding claims are compiled by three people in three Excel files and reach the owner a week late.

What we build

One dashboard refreshed every morning from the HMIS, Tally Prime and the attendance system, with a five-line WhatsApp summary to the owner and department heads.

Expected outcome

Typically, the owner stops asking for numbers and starts asking about causes, and the monthly review meeting runs from one screen.

How it works

1. The owner and the administrator choose the ten numbers that matter: OPD visits, new versus follow-up, bed occupancy, OT cases, average bill, collections by mode, TPA outstanding, pharmacy sales, staff attendance.
2. n8n pulls data from the HMIS, Tally Prime and the biometric attendance system each night into a single table.
3. A dashboard in Looker Studio or Metabase shows today, this month and the same month last year.
4. A short WhatsApp message with the five headline numbers goes to the owner at 8 am; department heads get their own numbers.
5. Occupancy below target, collections variance and claims pending beyond 30 days are highlighted automatically.
6. Each quarter we review with the owner which numbers are used and remove the rest.

Systems: HMIS · Tally Prime · biometric attendance system · Looker Studio or Metabase · WhatsApp Business API · n8n