Digital health · Clinical AI and patient engagement
Bot MD
A Singapore company that started as a chatbot answering doctors’ clinical questions and now sells AI agents running hospital patient engagement over WhatsApp and Viber. Its strongest evidence is a pandemic deployment a medical journal studied without ever naming the company.
Twelve thousand readings from a dormitory
In 2020, two migrant worker dormitories in Singapore with a combined population of 31,546 became the centre of the country’s largest COVID-19 outbreak. There was no on-site healthcare after hours. The residents spoke Bangla, Tamil, Mandarin and Hindi, and almost all owned a phone with WhatsApp on it.
Doctors from the National University Health System pushed a chat assistant onto those phones. Twice a day it asked each resident for a temperature, a heart rate and an oxygen saturation. The answers flowed into a dashboard that flagged anything abnormal.
The numbers were published in BMJ Open that December. Of 931 residents with COVID-19, 800 engaged, an 85.9 percent rate. They logged 12,511 discrete episodes of vital signs. Among 372 abnormal readings, 96 teleconsultations were started, seven of which went straight to emergency services and 18 of which pulled a physical review forward.
The paper calls the tool SGDormBot. It does not name Bot MD, and it states plainly that the chatbot developer took no part in the work. That is the cleanest way into this company: the best outcome data exists, it is real, and the company stands just outside the frame of it.
The problem it was built for
Dorothea Koh spent more than a decade selling medical technology into China, India, Indonesia and the Philippines for Baxter and Medtronic. What she kept noticing was not the machines. It was that doctors had almost no computers and always had phones.
“When you look at how much doctors have to do, it’s incredible. They have huge numbers of patients; they’re just overwhelmed,” she told TODAY in 2018.
Hospital knowledge sits in silos: protocols on an intranet, call rosters on paper, drug formularies elsewhere. Her design conclusion was a sentence she later gave Stanford Biodesign: “I figured it would be a lot easier to scale a solution that could be delivered through a mobile device.”
Two founders and a Coffee Bean
Koh, a biomedical engineer and a 2008 Stanford Biodesign fellow, started Bot MD in June 2018 with Yanchuan Sim, a natural language processing researcher formerly with Singapore’s Agency for Science, Technology and Research. The company dates the first sketches to nights and weekends in a Coffee Bean at Forum Galleria.
They got into Y Combinator’s Summer 2018 batch, worked out of a Palo Alto garage, and shipped version one in July 2018. Sim built the chat engine in house. “We do not use any chatbot technology from off-the-shelf providers,” Koh said in a 2021 investor interview.
By late 2018 about 1,000 doctors in 57 countries had downloaded it. It went viral in Egypt after one doctor posted about it. Then it stalled.
The turn that made it a business
Koh has been direct about why. The product was competing with a search engine, and losing. A cousin who is a doctor asked whether the bot could hold the on-call roster, the thing every resident copies by hand because the intranet is too slow. Bot MD added it and tested on 20 residents at National University Hospital in 2019. Within a month it was about 500.
Then came the pandemic, and two NUH doctors asking for something to monitor migrant workers at scale. That request became SGDormBot, and SGDormBot became Bot MD Care, the patient-facing product that now carries the company.
The commercial record since then is a list of hospitals. In Singapore: National University Health System, National University Cancer Institute, Tan Tock Seng, Changi General, Singapore General, Parkway Radiology, Alexandra Hospital, KK Women’s and Children’s. In the Philippines: the National Kidney and Transplant Institute, St. Luke’s, Philippine General Hospital. In Indonesia: Siloam Hospitals Group, 41 hospitals and 66 clinics treating more than five million patients a year, plus the Indonesian Medical Association and its 200,000 doctor members.
The positioning moved with it. The site today reads “Humans for care. AI for everything else.” and claims that “Bot MD AI Agents are deployed in live hospital and clinic environments”. The doctor chatbot that gave the company its name is no longer the headline.
What is proven, and what is still claimed
| Evidence | What the record shows | Source type |
|---|---|---|
| Peer-reviewed study | BMJ Open, December 2020. Two dormitories, combined population 31,546. Of 931 residents with COVID-19, 800 (85.9 percent) engaged; 12,511 vital sign episodes; 372 abnormal readings produced 96 teleconsultations, 7 emergency escalations and 18 earlier on-site reviews. Funded by Temasek Foundation. The tool is called SGDormBot. | Public record |
| Link from SGDormBot to Bot MD | The paper never names the company. The connection is made by a NUS Saw Swee Hock case study written in consultation with Koh, and by lead investor Monk’s Hill. | Company-informed |
| Other outcome figures | NUHS reported monitoring over 500 COVID patients at home and saving over 3,500 bed days, with a virtual ward review cited at 84 percent compliance. Post-operative outcome measure completion at KK Women’s and Children’s is reported at 78 and 63 percent. None appear in an indexed paper. | Company and investor stated |
| PubMed | A records pull on September 23, 2026 returned no papers indexed under the company or product name. | Not found |
| Singapore framework | HSA guidance on software medical devices (2020) and on clinical decision support (2022) sets risk classes by whether software informs, drives or makes a clinical decision. MOH and HSA jointly issue the Artificial Intelligence in Healthcare Guidelines, now version 2.0. | Public record |
| Device registration | No HSA registration or class assignment found as of September 23, 2026, and none published by the company. The products are described as scheduling, education, reminders and data collection, not diagnosis or treatment. | Not found |
| WHO recognition | Bot MD Care won the WHO Western Pacific Innovation Challenge in 2022. WHO attaches its own caveat: “no specific technology assessment was applied, nor any endorsement in practice should be implied”. | Independent |
| Named customers | Hospitals in three markets named by TechCrunch, The Business Times, KrASIA, Healthcare IT News and a 2025 investor memo. Siloam signed a memorandum of understanding in March 2023 covering outcome measures. | Independent |
| Security certification | ISO 27001, stated in the NUS case study. No certificate number or registrar published. | Company-informed |
| Scale | About 13,000 doctors in 2021, 18,000 by 2022, over 20,000 by 2023, over 50 institutions and 150,000 patients by September 2025. Every figure originates with the company. | Company-stated |
| Revenue | GetLatka publishes $2.3M for 2024 and labels it an estimate. No audited or company-stated figure exists. | Estimate only |
Read plainly: Bot MD has what most health AI startups lack, a long list of named hospitals that kept paying after the pilot. What it does not have is a published outcome study with its own name on it, a registered regulatory position, or a disclosed number a buyer could check. It sits on the administrative side of the clinical line, which is a large part of why it has moved this fast.
What to watch
- Whether the Siloam and KK Women’s and Children’s programmes produce a published paper.
- Any HSA classification. The moment an agent triages rather than collects, the regulatory question changes.
- The round behind Verge HealthTech Fund, whose September 2025 memo names no amount and no round.
- A disclosed revenue or customer count to replace the third party estimates standing in for both.
In their words
“When you look at how much doctors have to do, it’s incredible. They have huge numbers of patients; they’re just overwhelmed.”
Dorothea Koh, co-founder and CEO, TODAY, 2018 · Independent
“One of my personal goals in life is to reach 100 million patients by the age of 40.”
Dorothea Koh, EMPEA interview, 2021 · Interview
“The worst thing about what we were doing was competing with Google. What we should be doing is to take what Google doesn’t have, and deliver it in that way.”
Dorothea Koh, Her World Singapore, 2024 · Independent
“I joke that COVID-19 has been very good to us.”
Dorothea Koh, EMPEA interview, 2021 · Interview
“Doctors love Bot MD, and the high engagement numbers speak for themselves.”
Michele Daoud, partner, Monk’s Hill Ventures, 2021 · Investor
“Bot MD is a game changer for our post-op care at KKH.”
Dr Andrew Chou, orthopaedic surgeon, KK Women’s and Children’s Hospital · Customer, company site
“There was no direct involvement by the chatbot developer in the study or manuscript.”
Ko and colleagues, BMJ Open, December 2020 · Public record
Related companies
Sources
- Public recordRemote monitoring of marginalised populations affected by COVID-19
- Public recordDigital Health guidance, including AIHGle 2.0
- Public recordHSA software as a medical device and CDSS guidance
- Public recordRecords pull: no FDA, trial, PubMed or SEC hits
- IndependentBot MD Care, Innovation Challenge winner listing
- IndependentBot MD raises $5 million for expansion into Asian markets
- IndependentBot MD lands $5M for its AI clinical assistant
- IndependentBot MD raises US$5m led by Monk’s Hill Ventures
- IndependentMonk’s Hill leads Series A in Singaporean healthtech firm
- IndependentBot MD powers data collection in Indonesia
- Independent‘Medical Google’ for doctors?
- IndependentDoing the things Google can’t
- InterviewStartup Spotlight: Bot MD
- InterviewEmpowering doctors through AI assistance
- InvestorWhy We Invested in Bot MD
- Case studyBot MD Care case study
- CompanyBot MD site, about page and testimonials
- EstimateRevenue and headcount estimates
Profile researched and written by Healthcare Discovery. Last updated September 29, 2026.
