Team Builds Digital Product With AI Agents, Leaves One Human Behind
The startup’s office looks more like a server farm than a traditional workplace. Engineers, designers and even the project manager are code‑driven bots that chat, schedule meetings and push commits without ever logging in as a person. Only one member of the crew still types on a keyboard, and his daily agenda is to keep the swarm coordinated.
When the idea of a new offering emerged, the team faced an odd question: should the product mirror the composition of its creators, or should it aim for something else entirely? The answer, according to the lone human, hinges on the company’s core competence—building and deploying AI agents. He argues that a digital service built by agents for agents makes the most sense, especially when the rest of the staff already lives in that format.
The rationale behind that stance, and what the team hopes to achieve, unfolds in the following words.
So we figured if we're going to build a product, some kind of digital product, it should also include AI agents, since that's our area of expertise. Everyone is an AI agent except me, and I know a fair amount about AI agents. So we'll make a product that deploys AI agents to do something for you. But along the way, they don't usually use this phrase anymore, but they used to say, "The company eats its own dog food." It was a Google thing, Google uses Google products, I think.
Can a single human truly run a company of bots? Evan Ratliff’s HarumoAI puts that question into practice. The startup consists entirely of AI agents—employees, executives, even the product’s core logic—leaving Ratliff as the sole human overseer.
He says the goal is a digital product that “deploys AI agents to do something for you,” leveraging his own expertise in the field. While the architecture demonstrates that a team of agents can be assembled quickly, the article offers no data on user adoption, revenue generation, or long‑term maintenance costs. Moreover, the claim echoes Sam Altman’s vision of a billion‑dollar venture run by one person and many bots, yet the practical hurdles of accountability, error handling, and regulatory compliance remain unclear.
The experiment therefore serves more as a proof‑of‑concept than a finished business model. Whether such a structure can scale beyond a prototype, or attract investors without a traditional management layer, is still an open question. For now, HarumoAI stands as a test case in the ongoing exploration of AI‑centric organizations.
Further Reading
- Companies that replaced workers with AI are starting to rethink the tradeoffs - Futurism
- 2025: The year the frontier firm is born - Microsoft WorkLab
- AI agents will force companies to redefine performance and upskill their workers - Fortune
- Companies that signaled they are replacing workers with AI - Business Insider
- Future of work with AI agents - Stanford Human-Centered AI (HAI)
More in Business & Startups
Anthropic's Amodei says market will reward safe AI as 300,000+ use Claude
More than 300,000 startups, developers and companies have adopted a version of Anthropic’s Claude...
InnAIO T10 AI Translator Review: Feature‑Loaded with One‑Button Simplicity
Why does a translator that fits in the palm of your hand matter to a global workforce?
AWS unveils Kiro Powers with Stripe, Figma, Datadog for AI‑assisted coding
AWS rolled out Kiro Powers this week, pairing its new AI‑assisted coding engine with integrations...
Dario Amodei: Anthropic skips code reds, says enterprise AI differs from consumer models
Dario Amodei walked onto the stage with a clear agenda: to draw a line between the AI tools...
Healthify and Novo Nordisk India debut AI-driven obesity support program with Ria
Why does a pharma‑tech partnership matter now? In India, obesity is climbing faster than many...
Common Questions Answered
How does HarumoAI’s team composition differ from that of a typical startup?
HarumoAI’s workforce is made up almost entirely of code‑driven AI agents that handle engineering, design, and project management tasks. Only one person, Evan Ratliff, remains as a human overseer to keep the swarm coordinated.
What specific responsibilities does Evan Ratliff hold as the sole human at HarumoAI?
Ratliff acts as the single human overseer, managing the daily agenda of the AI swarm, ensuring bots stay aligned, and making high‑level decisions that the agents cannot handle autonomously.
What is the main goal of the digital product HarumoAI intends to release?
The product is designed to deploy AI agents on behalf of users, leveraging the company’s expertise in agent‑based automation to perform tasks without human intervention. It reflects the team’s belief that the product should mirror its AI‑centric composition.
How does the article relate HarumoAI’s approach to the concept of “eating its own dog food”?
The founders note that, because every employee except Ratliff is an AI agent, they are naturally building a product that uses the same technology they employ internally—mirroring the Google practice of using its own products to validate them.