Dorothea Koh built Bot MD in emerging-market hospitals where the answer a doctor needed was rarely one search away. Today her AI agents field the messages, bookings and follow-ups that used to jam a hospital's phone lines.
There is a specific kind of silence that happens in a hospital at 2am. A junior doctor stares at a drug they have not prescribed in months, unsure of the renal-adjusted dose. The senior on call is asleep. The reference book is a corridor away, and the search engine on the phone returns a forum thread from 2011. That gap - between a question and the right answer - is where Bot MD was born.
The company, based in Singapore and founded in 2018 by Dorothea Koh and Yan Chuan Sim, started as a deceptively simple idea: an AI chat assistant that a doctor could message like a colleague. Ask it about a drug, a protocol, a guideline, an on-call roster - and it answers, on a phone, in seconds, using the hospital's own content rather than the open internet. Koh had spent years running healthcare operations for Baxter and Medtronic across Indonesia, Singapore and the Philippines, and she had watched clinicians in busy, under-resourced hospitals improvise their way to information. Bot MD was her attempt to close that loop.
The first version of Bot MD pointed inward, at clinicians. It could be trained on any hospital's formularies, protocols, clinical guidelines, call schedules and even radiology reports, then sit in a doctor's pocket as a searchable assistant. During the pandemic, when guidance changed weekly and staff were stretched, that turned out to be useful at scale. The app spread by word of mouth through hospital wards across Asia.
But Koh and Sim noticed the same bottleneck was choking the other side of the counter. Patients were flooding hospitals with messages - where is my report, can I move my appointment, what am I supposed to do after discharge - and the phone lines and front desks could not keep up. So Bot MD widened its aim. Today its headline product is an AI patient-engagement platform: a suite of agents that automate the routine, high-volume conversations a hospital has with the people it treats.
AI shouldn't replace care teams - it should give them leverage.Bot MD's operating principle
That sentence is doing a lot of work. Plenty of healthcare AI companies have tried to play doctor, promising diagnosis from a chat window. Bot MD deliberately does not. Its agents handle logistics, education and follow-up - the unglamorous connective tissue of care - and hand off to a human the moment judgement is required. The company sums up its own boundary in four words that appear across its materials: "Humans for care. AI for everything else."
Rather than one all-purpose chatbot, Bot MD ships a set of specialised agents, each aimed at a specific job a hospital would otherwise do by hand. Hospitals can compose them into a multi-agent suite that walks a patient from first enquiry through booking and all the way to post-discharge check-ins.
The survey agent is quietly one of the most interesting. It collects PROMs, PREMs and CSAT data - the patient-reported outcome and experience measures that hospitals are increasingly judged on, but that they have rarely had the staff to gather consistently. A monitoring agent runs post-discharge check-ins, flagging the patients who need a callback. A pre-admission agent collects intake details and consent before anyone arrives. None of it is dramatic. All of it is the work that, left undone, turns into a missed appointment or a readmission.
One of Bot MD's least obvious but most consequential choices is where it meets patients. It does not ask them to download an app. Its agents operate over the messaging channels people already live in.
This is a direct inheritance from Koh's years working in Indonesian and Filipino hospitals, where WhatsApp is not a convenience but the default operating system of everyday communication. A patient-engagement product that required a separate app would have quietly excluded the very people Bot MD was designed to reach. Building on chat channels was the difference between a demo and a deployment.
Bot MD has helped us automate radiology scheduling and significantly reduce our call center volume.Yujuan Tan, CEO, Parkway Radiology
In an era when anyone can wrap a language model in a chat box, Bot MD's defensibility does not come from the model. It comes from what the agents are plugged into. The platform connects to a hospital's EMR, HIS, CRM, scheduling systems, queue management, content management and payment gateways, then answers using that specific institution's data. A generic assistant can tell a patient what a colonoscopy is. Bot MD's can tell them when their slot is, what to do the night before, and reschedule it - because it is wired into the systems that actually hold those answers.
Two more things make this saleable to risk-averse healthcare buyers. First, speed: Bot MD says it can deploy agents within roughly a week and start driving patient uptake quickly. Second, trust: every AI interaction is securely logged and auditable, and the company holds ISO 27001, ISO 27017, ISO 27018 and SOC 2 certifications. In a domain where patient safety and privacy are non-negotiable, an audit trail is not a feature - it is the price of admission.
Bot MD sells to hospitals, clinics and healthcare groups as a B2B SaaS platform, with deployments spanning Singapore, Malaysia, Indonesia, the Philippines and Hong Kong, across specialties from radiology and oncology to renal care, paediatrics and chronic-disease management. Its named users read like a roster of the region's larger institutions: Parkway, National University Hospital, Singapore General Hospital, KK Women's & Children's Hospital and National Cancer Centre Singapore, with St. Luke's Medical Center and Philippine General Hospital among Philippine deployments.
The economics are framed around two levers a hospital administrator understands immediately: fewer calls the front desk has to field, and more appointments that actually get booked and kept. Bot MD pitches its impact in terms of reduced call-center load and increased revenue from conversations that previously went unanswered. It is an unusually concrete value proposition for an AI product - less "transformation," more "the phones stop ringing off the hook."
Dorothea "Dot" Koh graduated summa cum laude from Northwestern with a degree in biomedical engineering and economics, holds a master's in bioengineering from Stanford, and before founding Bot MD was general manager of Baxter across Indonesia, Singapore and the Philippines. That operating background - not a lab, but the messy reality of running healthcare businesses in emerging markets - is stamped on every product decision, from the channel strategy to the refusal to diagnose. Her co-founder, Yan Chuan "YC" Sim, is a PhD researcher in natural-language processing and machine learning who architects the platform and its integrations.
Bot MD began, fittingly, as a weekend project before it became a Y Combinator company. In February 2021 it closed a US$5M Series A led by Monk's Hill Ventures, with SeaX, XA Network, SGInnovate and Silicon Valley figure Steve Blank participating, earmarked for expansion across Indonesia, the Philippines, Malaysia and India. In 2022 it was named among the winners of the World Health Organization's Western Pacific Innovation Challenge.
Globally, Bot MD lands in the same conversation as patient-engagement and healthcare conversational-AI vendors such as Hyro, Notable Health, Klara and Luma Health, while its clinical-reference roots put it adjacent to tools like UpToDate and plain medical search. What separates it is less any single feature than a posture: it is built for emerging-market realities first, runs on the channels patients already use, and draws a hard line at diagnosis. That focus is also its risk - Bot MD is a roughly 29-person company competing in a category that global players are racing into. But it has something the newcomers cannot buy quickly: years of live deployments inside real hospitals, and the integrations to prove it.
The bet underneath all of it is modest and, for that reason, credible. Bot MD is not trying to replace the doctor at 2am. It is trying to make sure that by the time a patient reaches that doctor, every message has been answered, every appointment booked, every follow-up sent - so the human hours go to the part of medicine only humans can do.