The One-Person Company: How Solo Founders Spend Their Time with AI Agents
The headlines are about the one-person unicorn. The reality, for most solo founders, is a Tuesday that starts with three agent runs to review, a support email, a half-finished landing page and a sales call nobody else is going to take.
The one-person company is no longer a niche idea. More founders are starting alone, and AI agents now handle work that used to need a first hire. This post looks at the data behind the trend, what agents actually do in a solo business, and how a solo founder's week changes when part of the work is delegated to software.
What Is a One-Person Company?
A one-person company is a business with a single owner-operator and no full-time employees. It can still use contractors, freelancers and software. What makes it "one-person" is that one person makes the decisions and carries the work that isn't outsourced.
The term overlaps with a few others:
- Solo founder: one founder, often aiming to raise money and hire later.
- Solopreneur: one owner, usually planning to stay small and profitable.
- Freelancer: sells their own time to clients rather than a product.
In 2026 the difference is mostly about ambition. The day-to-day looks similar: one calendar, many roles, and a growing list of AI tools.
How Many Founders Are Going Solo?
More than ever, by the available data. The Carta Solo Founders Report, based on US startups that use Carta, found that the share of new startups with a solo founder rose from 23.7% in 2019 to 36.3% in the first half of 2025.
Money hasn't caught up yet. Solo-led companies made up 30% of startups founded in 2024 but received only 14.7% of cash raised in priced equity rounds that year. Solo founders also hire their first employee sooner: a median of 399 days after incorporation, against 480 for multi-founder companies.
Developers show the same shift. In the 2026 Stack Overflow Developer Survey, the share of respondents working in an organization of one jumped from 4% to 10%. Stack Overflow's own reading is that some of them may be using AI agents to run a viable company alone. Part of that jump likely reflects a tough hiring market too. If companies won't hire you, starting your own is the obvious alternative, and AI has made the first version cheaper to build.
Has Anyone Built a Billion-Dollar Company Alone?
Sort of. In 2024, Sam Altman predicted that AI would make a one-person billion-dollar company possible. In April 2026, The New York Times reported on Medvi, a telehealth company that sells GLP-1 weight-loss drugs online. Its founder, Matthew Gallagher, set it up in two months with about $20,000 and a stack of AI tools for code, ads and customer service.
According to the report, Medvi brought in $401 million in its first year and was on pace for $1.8 billion in sales. It's a remarkable story, with two footnotes. Gallagher's brother joined as the only full-time hire, and doctors, prescriptions and shipping were outsourced to other companies, with contract engineers and account managers on top.
That's the honest shape of most one-person companies. One person, plenty of help, and the help is a mix of contractors, platforms and AI.
What Do AI Agents Actually Do in a One-Person Company?
Less than the hype and more than nothing. A FreshBooks survey of 500 solopreneurs and microbusiness owners, run by Wakefield Research, found:
- 86% try AI before hiring someone for a task they can't do alone.
- 90% say AI makes it easier for one person to start and run a business.
- Marketing and design and customer service are the most common uses, ahead of admin work such as finances.
For technical founders, the agent is usually a coding agent. Among Stack Overflow's 2026 respondents who use AI, Claude Code (66%) and GitHub Copilot (59%) were the most-used agents, and about a third of daily AI users spent four or more hours a day with AI tools.
So agents are now part of the default stack. The open question is what that does to the founder's own time.
How Do Solo Founders Spend Their Time with AI Agents?
There's no large time-use study of one-person companies yet, so treat this section as a framework, not data. Based on the surveys above, a solo week tends to split into five buckets. Your mix will differ by business, but the buckets rarely change:
| Bucket | What agents can take | What still needs you |
|---|---|---|
| Building | First drafts of code, tests, refactors | Product decisions, architecture, final review |
| Selling and marketing | Drafts of posts, ads, outreach, landing copy | Talking to customers, positioning, closing |
| Support | Triage, first replies, docs lookups | Edge cases, angry customers, refunds |
| Admin | Categorising expenses, drafting invoices | Approving, filing, deciding |
| Supervising agents | Nothing | Writing tasks, reviewing output, fixing mistakes |
Admin is the bucket solo founders underestimate most. In the FreshBooks survey, 57% still spent more than 10 hours a week on admin work like invoicing, bookkeeping and taxes, even though finance trailed other areas in AI use.
The last row is the new one. A year or two ago, it barely existed. Today it can quietly take an hour or more of a solo founder's day: writing clear instructions, waiting for runs to finish, reading diffs and drafts, and redoing the parts that missed. It's real work, but it rarely shows up on a to-do list or in a plan for the week.
Is Supervising AI Agents the New Hidden Job?
For many solo founders, yes. The same FreshBooks survey found that 65% said that while AI had improved productivity, tasks like reviewing its work and checking for accuracy create a risk of "AI burnout".
Reviewing AI output is real work, even when it doesn't feel like it. Photo: Felicity Tai on Pexels
That fits a wider pattern. In METR's 2025 study, experienced open-source developers using AI tools took 19% longer to finish tasks, while believing they had been faster. Some of that time went into prompting, waiting and checking.
None of this means agents don't pay off. It means their cost is easy to miss. Reviewing an agent's pull request doesn't feel like work in the moment, but an hour of it every morning is five hours a week. If you measure your solo business only by output, this cost stays invisible. We covered which numbers are worth tracking in productivity metrics for solo founders.
How Should You Plan a Solo Week Around AI Agents?
Treat agents like contractors who work fast and never ask questions. That suggests a few habits:
A fixed weekly rhythm keeps agent work from taking over the day. Photo: Tara Winstead on Pexels
- Batch agent hand-offs. Write tasks for agents at set points in the day, not every time a thought appears. Constant hand-offs fragment your focus.
- Batch reviews too. Review agent output in one or two blocks rather than as it arrives.
- Protect one maker block a day with no agent runs, chat or email. Product decisions need uninterrupted time.
- Do one customer-facing thing every day. Building feels productive, especially with an agent shipping code. Selling is what keeps a one-person company alive. If you need low-cost ideas, see our growth hacks for solo founders.
- Keep admin in a fixed weekly slot. It expands to fill any time you leave open.
- Review the week on Friday. Compare where you meant to spend time with where it went.
The goal isn't to squeeze more output from agents. It's to make sure the work only you can do still gets done.
How Do You Know Where Your Week Actually Went?
This is the hard part of being a company of one. Nobody else sees your calendar, and memory is generous. Most solo founders believe they spend more time building and selling than they do.
Manual timers rarely survive a week of switching between an agent, a customer email and a deploy. Didon does the tracking automatically. It runs in the background on your Mac, takes a screenshot every 30 seconds and analyzes it on-device with a local model, so nothing leaves your machine. You get a daily journal with time per project and a short summary, and you can export the data as CSV if you want to run your own numbers.
Because it reads what's on screen, time spent reviewing an agent's work lands under the project it belongs to. After a couple of weeks, you can see the real split between building, selling, support, admin and supervising, rather than the one you'd guess.
What Should You Take Away About the One-Person Company?
- Solo-founded startups rose from 23.7% to 36.3% of new companies between 2019 and the first half of 2025, per Carta.
- Developers working in an organization of one went from 4% to 10% in a year.
- The billion-dollar solo story is real but comes with contractors, platforms and a second person.
- AI agents take first drafts in code, marketing and support. They add a new job: supervising them.
- The founders who do well alone will be the ones who know where their own hours go.
One person can now do what used to take a small team. That still adds up to one person's week, and it's worth knowing how you spend it.
Related Posts
The Future of Work in 10 Years: What Work Could Look Like by 2036
The future of work in 10 years, grounded in WEF, McKinsey, Stanford and Nature data: jobs, AI agents, remote work, shorter weeks and the skills that will matter.
How Do Developers Work in 2026? What the Data Says About the Developer Workday
How developers work in 2026, from the latest Stack Overflow, JetBrains, METR and Microsoft data: where they work, how much they code, and what AI agents changed.
