Five steps, about a minute: details, doctor, day, time, confirmation with an appointment ID.
WhatsApp Booking Agent
Patients book on the number a clinic already advertises, at any hour, in the language they normally write. Only genuinely free slots are offered, and anything clinical goes to a person straight away.
The problem
What it was like before
A clinic receptionist costs a salary and works office hours. Calls arriving outside those hours are appointments lost, and calls arriving during them interrupt whoever is looking after the patient in the room.
Every alternative asks the patient to install something. For a single appointment, they will not.
How it looks
Five steps, about a minute
The booking form as a patient sees it, cycling through its steps.
✕Appointment book karein
What we built
What it does
Patients who would rather just type can do that instead.
Slots are checked against the doctor's sessions, blocked days and every appointment already taken, so a double booking cannot be offered.
Roman Urdu, Urdu or English, including switching mid-sentence.
Written to the clinic's own sheet rather than a new system to learn.
Urgent messages get 1122 guidance at once and alert staff. Medical advice is never given, and that is not configurable.
Complaints, late changes and anything unusual go to a person with the thread attached.
Where it got to
In production
24/7Including nights and holidays
About a minuteFrom first message to appointment ID
No appNothing for a patient to install
Our view
Distribution beats features
Most clinic software fails on distribution rather than capability. The booking screen is not the hard part; getting a patient to arrive at it is.
WhatsApp is already open on the phone, which is why the whole product lives inside it and why there is no portal, no login and no app.
Built with
Tech stack
- WhatsApp Business API
- WhatsApp Flows
- Google Sheets
- PostgreSQL
- Docker
- Caddy
The work behind it
Delivered under WhatsApp AI Assistants.
More work
Other things we have built.
Sanctions Screening Agent
Watchlist alerts arrive all day and most are false matches. The agent clears the obvious ones against a rule engine backed by an LLM and sends every disagreement to a person with the evidence already assembled.
Read itLaw-Enforcement Request Intake
Requests arrive as letters with scanned attachments. The agent reads the mailbox, OCRs the attachments, verifies identifiers against core banking, classifies the intent and routes it — with an audit trail behind every step.
Read itInternal Policy Assistant
Staff ask a policy question in plain language and get an answer drawn strictly from approved documents, with the source named. Nothing leaves the bank, and the assistant says so when the documents do not answer.
Read itCTR to goAML Conversion
A compliance team spent hours a week turning Currency Transaction Reports into goAML XML by hand. Parser, XML builder, schema validator and audit trail, behind a interface officers actually use.
Read itRegulatory Reporting Engine
A registry of every periodic return a bank owes. The engine pulls the data, builds each report against the mandated template, then schedules and submits it behind a maker-checker step.
Read itAutonomous Social Media Agent
A ReAct loop that researches, reads its own history to avoid repeating itself, writes a strategy memo, critiques its own drafts, publishes, and learns from every rejection.
Read itMulti-Agent Marketing System
Two agents on one platform with a person approving everything before it ships. One researches and takes a position; the other plans a week of posts reviewed as a batch.
Read itHave something like this?
Describe the process and we will come back with whether it is worth automating, roughly what it would take, and what we would build first — or tell you plainly if it is not a job for us.