Medvi, two people and an army of AI agents
I just watched Raymond Chin talk about Medvi. Rp 6.8 trillion in revenue in 2025, and on track for Rp 30 trillion this year.
Its competitors need 2,400 employees to hit numbers like that. Medvi? 2 people. The rest is AI.
Using AI vs employing it
What makes Medvi different isn’t that the founder is a technical genius.
Matthew Gallagher isn’t ex-Google, isn’t an engineer, isn’t from Silicon Valley. Gallagher grew up in a trailer park, has a background in digital marketing, and learned AI from YouTube and ordinary browsing.
All Gallagher had was an understanding of marketing + a lot of curiosity + the willingness to use AI to a level other people can’t be bothered to reach.
And “using AI” here isn’t our version of it.
Us: open ChatGPT, ask something, close it.
Gallagher: build an army of AI agents that work 24 hours a day.
There’s a marketing agent that checks every morning which ads are working, then makes new creatives, uploads them, monitors them and repeats. There’s a customer service agent that uses ElevenLabs. A finance agent. An analytics agent. They all talk to each other.
Gallagher doesn’t use AI. Gallagher employs it.
This made me rethink how I use Claude and ChatGPT myself.
So far I’m still at the “tools” level: open it when I need it, close it when I’m done. Raymond says 90% of Indonesians are still at this level. But there’s a level above it: AI that has its own schedule, its own access to data, and works while we sleep.
The difference is like having a calculator vs having an assistant.
The dark side
But Raymond is also honest about the dark side, and this part matters.
Medvi used fake doctor videos in its ads, and AI-generated before-and-after photos of patients. It’s already received a warning letter from the FDA. The growth is fast, but the methods are very gray.
What I take from it: the strategy is worth studying, but don’t copy the ethics. AI is leverage. It multiplies anything, including the bad stuff.
Leverage multiplies zero too
What stuck with me most was a quote from Palantir’s CEO that Raymond brought up.
Leverage is multiplication. If your output is 10, with AI it can become 1,000. But if your output is 0 because your skills aren’t relevant or you haven’t learned the tools, multiply it by anything and it’s still 0.
Raymond’s point is that the layoffs aren’t just AI taking jobs: the people most exposed are the ones who haven’t learned to work with it.
The World Economic Forum says that by 2030, 92 million jobs will disappear and 170 million new ones will appear.
The numbers are actually good. But the new jobs will likely go to people who can work with AI the way Matthew does, even at a much smaller scale.
What I want to try
Three things I want to try myself after watching this:
- Stop thinking of AI as a chat helper. Start thinking about which repetitive work could become an agent.
- Level up my prompting: give it a role, context and an objective, and let the AI ask questions back.
- Hold on to ethics. Growth without trust doesn’t last.
I’m still at step 1 myself. But I don’t want to stay there long.
One line from the video is still going around in my head. A Rp 30 trillion company run by 2 people used to sound like a theory. Now there’s an actual person, actual numbers and an actual FDA warning letter.
The question isn’t whether this is possible anymore. It’s whether I learn to use it.
Watch the full video, “Perusahaan 29 Triliun Dengan NOL KARYAWAN” (roughly, “A 29 trillion company with zero employees”), on Raymond Chin’s YouTube channel (in Indonesian). It’s really worth watching.