You've left your job and shipped your first product alone, and in that first week you stare at your to-do list and go briefly numb. The core feature took only a few days. But beneath it stretches a queue: landing-page copy, payment integration, listing on distribution channels, business registration, customer support. Only now does it sink in, viscerally, that a design team, a legal team, and an accounting team used to handle all of this for you. One night you spend two hours with three windows open on your monitor. One is a quote from a freelance developer. One is the checkout screen for an AI design tool. One is an online course. The night wears on and you still haven't decided whether to hand off a single payment integration, wire it up yourself with a tool, or learn to do it.
What's stuck isn't the integration itself. Integration can be learned; quotes can be compared. What you don't have is a rule for deciding who should do what.
For a solo founder, the first problem isn't revenue — it's allocation
For most people who start a company alone, the first problem they hit — before worrying about revenue — is allocation. Dozens of times a day, the same question comes back around: do I do this myself, or hand it off? Decide it case by case, with no rule, and you typically collapse in one of two ways. You do everything yourself until you run out of time, or you outsource everything until quality and cash run out together. Many solo founders spend their first three months in the first mode. Teaching themselves payment integration and building settlement spreadsheets, they let product improvements slip, and one month after another they do the least of the very thing they do best.
So the question has to be simple. Start by deciding where to draw the boundary around your own work. The delegation line is what you get when you lay out every task in your business and draw a line between the work you'll do yourself and the work you'll hand off. A task is the smallest unit of work — a job function divided until it can't be divided further. “Marketing” is a function; “drafting the weekly newsletter” is a task. The line has to be drawn finer than the function, at the level of tasks. Within that same marketing function, drafting might be work to hand off while approving what goes out stays with you.
The question of why companies exist lands on your desk
In 1937, the economist Ronald Coase asked a strange question in “The Nature of the Firm.” If the market is so efficient, why do people build companies and work inside them? His answer was transaction costs — the costs of finding a counterparty in the market, negotiating terms, and monitoring the result. When buying something on the outside costs more than handling it inside, a firm brings the work in-house; when the reverse is true, it sends the work out. The boundary of the firm, Coase argued, settles at the point where the two costs are equal.
That night you spent was a transaction-cost night, too. Finding developers and comparing portfolios is search cost; negotiating the quote and the number of revisions is bargaining cost; checking the deliverable and sending it back for fixes is monitoring cost. Even if the freelancer's fee is 300,000 won (about US$220), add these three costs and the real price climbs higher. Reframe your hesitation as a transaction cost, and that night turns into a problem you can handle with arithmetic.
For a long time, this theory was used only to explain mergers and outsourcing strategy at large corporations. A solo founder, unable to hire employees, simply had no “bring it in-house” option to begin with. The arrival of AI tools broke that premise. With no search and no negotiation, for a few tens of thousands of won a month (a few dozen dollars), a processing capacity that keeps running at three in the morning appeared — and suddenly even someone working alone had a domain they could call “inside.” The outside widened, too. As Arun Sundararajan laid out in his work on the gig economy, digital platforms opened a market for renting outside labor one task at a time, without hiring anyone. With AI on the inside and a task-by-task outsourcing market on the outside, the century-old question of why firms exist comes down to this: What do I do, what does AI do, and what gets outsourced?
The delegation rule that usually comes to mind here is: hand off what you can't do. It's intuitive, but only half right. This is exactly what David Ricardo showed with the principle of comparative advantage in 1817. Comparative advantage holds that, regardless of absolute skill, both parties gain when each specializes where its opportunity cost is lower. Picture a developer whose code is better than an AI's. In absolute skill, the human wins. But every hour he spends coding is also an hour he could have spent on product planning or customer interviews. If an hour of planning creates more value than the gap in code quality, then handing the coding to the AI and keeping only the final review grows the output of the business as a whole. The test isn't “is the AI better than me?” but “what am I giving up while I hold on to this?” The tasks we grip precisely because we're good at them are often the ones whose surrendered value is highest.
And this boundary isn't a line you draw once and leave. By Coase's logic, the line should move every time transaction costs change — and the capability and price of AI tools shift on a scale of months. Subtitling you outsourced last year becomes something you wire up yourself with a tool this year; the data cleanup you used to do yourself slides over into the AI column. For a solo founder, the delegation line is a problem you re-solve every quarter. The finer you've drawn a moving line, the easier it is to redraw.
Three moves for drawing the line
This hesitation shrinks down to a single table. Three moves are enough.
First, write down twenty tasks from the past two weeks, each as a verb. The trick is not to write them big, like “marketing,” but to split them into units built around a verb — “write an Instagram post,” “confirm a client's payment.” If your memory is hazy, scroll back through your calendar and your transaction history. When delegation feels overwhelming, it's usually because the question is bundled at the size of a job function; the moment you break it down to the size of tasks, half of them are already decided.
Second, choose what to keep before you choose what to hand off. Tasks where judgment compounds into an asset — talking with customers, setting prices, making the final call on quality — are yours. The output of these tasks isn't just that day's deliverable. The more you repeat them, the more a feel for your customers and the data about them accumulate on your side, and that accumulation is hard for anyone else to copy. Choose what to hand off first, by contrast, and you typically start with the work you least want to do — and inside the work you least want is often a task where judgment compounds, like customer support. One founder who went solo making course videos outsourced everything from the first month — editing, subtitles, even replying to comments — and three months later was left unable to explain why any given video on his own channel had caught on. He'd handed off the tasks that needed judgment before any judgment had a chance to accumulate.
Third, split the remaining tasks between AI and outsourcing — but put the cost of delegating into the math. Tasks whose steps you can write down as a document and whose mistakes you can reverse belong to AI. One caveat: a task you've never done yourself is hard to review even after you hand it off, so do it yourself once, until you have a standard for what good output looks like. Tasks that carry legal liability, involve physical work, or require expert judgment whose steps can't be transferred belong to outsourcing. On each row of the outsourcing column, write your own time in parentheses next to the fee — the time it takes to get the quote, explain the work, and check the result. As Michael Jensen and William Meckling set out in agency theory, the party who delegates and the party who takes on the work have different interests and asymmetric information, so monitoring costs come attached. Multiply that time by your hourly rate, add it to the fee, and that sum is the task's real price.
From productivity to profitability
Once you've drawn the line, time is left over. But time saved through delegation is, on its own, only a savings asset. It's worth no more than its hourly value times the hours saved, and it has a ceiling. For saved time to turn into revenue, the value of that time has to be set by the tasks you kept inside the line — by judgment and by relationships. Drawing the line looks like a productivity problem, but the moment you decide what to keep, it turns into a profitability problem — a decision about what you'll earn from.
There's a byproduct, too. Handing a task to AI requires writing the steps down as a document and defining a standard for review — and that bundle of instructions and review procedure is both a savings asset you make once and reuse forever and a latent asset you could externalize to other solo founders doing the same work. The delegation rule turns out to be the first place where “make it once, use it twice” starts working from month one. So before you draw up a revenue plan, fill in the task table this week. And put the next quarter's first-week refill on your calendar now, too. In a company with no employees, the trail left by a moving boundary serves as the org chart.
This series grows out of a single manuscript that re-binds the standard theories of accounting, economics, management, and investing — across disciplinary lines — into the problems of running a business alone. The delegation line is the starting point of that map. The next installment looks at how the judgment tasks you kept inside the line build into a moat no one can copy, and why that asset gets recorded as zero won on the accounting books. If you've grown comfortable with building but have stalled at the question of what to keep, the next installment is where the real argument begins.
Concept appendix
- Transaction cost theory — Introduced by the economist Ronald Coase (1937) in “The Nature of the Firm.” It refers to the cost of finding a counterparty in the market, negotiating, and monitoring, and explains that a firm brings work in-house when that cost exceeds the cost of handling it internally. Oliver Williamson added the conditions of asset specificity and uncertainty. As AI lowers the cost of internal handling, the same logic applies to how one person allocates work. - Comparative advantage — Put forward by David Ricardo (1817) to explain trade between nations. The principle holds that, regardless of absolute skill, both parties to a trade gain when each specializes where its opportunity cost is lower. Only the trading partner has changed — from a country to an AI — and it applies, unchanged, to how you divide tasks between yourself and an AI. - Agency theory — Set out by Michael Jensen and William Meckling. The view holds that because the party who delegates and the party who takes on the work have different interests and asymmetric information, monitoring costs and residual loss come attached. Compare only the outsourcing quote against your hourly rate and these costs drop out of the math.
About the series — An insight series based on the manuscript “Running a Company by Yourself: A Management Framework for the Solo Founder in the Age of AI.” Each installment takes on a single management decision.
