Fractional AI vs hiring, the math, function by function.

Every funded team under fifty hits the same fork in the road around month nine. The seed round is in the bank. The first ten hires shipped a product that works. Revenue is small but real. The board asks the question that always comes next, which is what the org chart looks like in twelve months. The founder opens a spreadsheet, lists the gaps, and starts mapping roles to budget. Two SDRs. A content marketer. A finance analyst. A support manager. Maybe a head of growth if the runway holds. Eight to ten hires across the next eighteen months. The plan looks reasonable on the slide.

The plan is not reasonable on the cash flow model. Two SDRs loaded is one hundred and sixty thousand a year before tools and management. A content marketer plus a freelance writer is one twenty plus six thousand a month in agency overflow. A finance analyst is ninety-five loaded with thirty thousand more in dashboard licenses sitting in the stack unused. A support manager plus an overnight BPO is one twenty to two hundred depending on the contract. Add the head of growth, the recruiter fees, the equity refresh pool, the laptop budget, the seat licenses. Year one loaded cost for a four-person team across functions is four hundred to eight hundred thousand. That is before any of them ramp to full output.

None of those hires ship anything in month one. Two months of recruiting, two months of onboarding, three to six months of ramp before the SDR is hitting quota or the marketer is shipping the cadence. Year one is mostly the cost of waiting. By the time the team is producing, the next fundraise is on the horizon and the runway model assumes another round of hires to scale what just started working. The org chart compounds. The output ceiling per human stays flat.

The other path is to take one of those four functions, run it as a fractional AI department, see the output inside fourteen days, and decide whether to repeat the move on the next function before you sign the next offer letter. The math is not abstract. It is a single retainer per function, smaller than the loaded cost of one of the hires it replaces, with output live in two weeks instead of six months. We covered the underlying shape in What is a Fractional AI Department. This page is the head-to-head, function by function, with the numbers in plain view.

Recruiting an SDR takes sixty days on the conservative end. Two weeks to write the JD and post it, three weeks of screening and first-rounds, two weeks of finals and reference checks, two weeks of notice from the candidate. That is before they sit down. Once they sit down, onboarding is two weeks for tools and product, and ramp to quota is another three to six months. Best case from the moment you decide to hire to the moment the rep produces a qualified opportunity is one hundred and twenty days. Realistic case is closer to one hundred and eighty.

A content marketer is the same shape with a different curve. Sixty days to recruit, two weeks of context loading, six to eight weeks before the brand voice is internalized enough to ship long-form pieces that do not need rewriting. Three months in, you have one or two articles live and a marketer who is starting to understand the product. A finance analyst is faster on tools and slower on context, because the context is your entire ops stack and the calls the COO already made about what the numbers mean. Three to six months before the analyst is producing reports your board trusts without the COO doing the reconciliation pass.

During those months you are paying the loaded cost and producing nothing. Fifteen thousand a month for an SDR you cannot send into the queue. Eight thousand a month for a marketer who is reading old decks. Twelve thousand a month for an analyst who is still learning the Stripe schema. Across four hires the ramp tax is one to two hundred thousand of compensation against zero output. The compounding part is what the rest of the team did not do while the new hire was learning, because the senior people were the ones doing the onboarding.

A fractional AI Sales Department is live in fourteen days. Day fifteen, five hundred personalized touches are in the queue and the first warm replies are landing. The AI Content Department ships its first long-form piece by day fourteen and hits weekly cadence by week four. The AI Ops Department runs its first board update at week three with the COO editing the draft in twenty minutes. The AI Support Department is deflecting tier-one in seconds by week two. There is no ramp tax. The first invoice and the first output land in the same month.

The ceiling on human output per function is not a soft constraint. It is physics. An SDR who is researching the prospect, writing the email, sending it, logging it in the CRM, and moving to the next one cannot ship more than 80 to 100 touches a day before quality collapses. A content marketer who is researching the topic, writing the draft, editing, formatting, and publishing cannot ship more than one to two long-form pieces a week, and that is on a good week with no meetings. A support rep can resolve 30 to 50 tier-one tickets a day before the replies get curt. A finance analyst can answer ten to twenty ad-hoc questions a week before the queue backs up.

You can stack humans to lift the ceiling, but the cost is linear and the output ceiling per dollar barely moves. Two SDRs produce 160 emails a day for 160 thousand of loaded cost. Three SDRs produce 240 emails a day for 240 thousand. The ratio is constant. The compounding loss is what the senior people did not do while they were managing the SDR team, because the manager-of-SDR layer is its own salary the moment you hit three reps.

Fractional AI removes the per-unit labor cost. The same agent runs five hundred touches a day or fifty, eight articles a month or twelve, fifty board questions or five hundred. The retainer does not move with the volume because the volume is not human-bottlenecked. That is the only reason the year-one numbers on a fractional content department come in below the cost of a single in-house marketer while shipping six times the output. You stopped paying for the labor and started paying for the function.

The implication for the org chart is direct. The hires you were planning to make to scale output are mostly not hires anymore. They are scope expansions on the existing retainer. The hires you still want to make are the ones where human judgment is the work, not the throughput. Founders, senior operators, CFO, head of product, head of sales. The middle layer of execution roles compresses out of the plan, and the senior layer gets more leverage per dollar.

The hiring model is per-person pricing. Every additional unit of output requires an additional person, an additional salary, an additional manager, an additional desk, an additional set of tools, an additional layer of HR overhead. Output scales with headcount, headcount scales with cost, and the cost line on the P&L moves in lockstep with the FTE count. You can read your year-one burn directly off the org chart. That is the whole shape of the model.

The fractional model is per-function pricing. You buy the function. The function ships the output. Inside the function, throughput scales without the cost line moving, because the throughput is not human-bottlenecked. A retainer for the support department covers one ticket a day or a thousand. A retainer for the sales department covers fifty touches a day or five hundred. The variable that used to be labor is now machine time, and machine time priced as a department is fixed.

The implication for the cash flow model is that the year-one math is not a per-month subtraction. It is a different curve entirely. Hiring four functions to a serviceable level across eighteen months runs four hundred to eight hundred thousand fully loaded, and you are still managing four sets of humans, four ramp cycles, four burnout windows. Running four fractional departments runs four monthly retainers, none of which exceeds the loaded cost of a single hire in that function, with output live in fourteen days and scope flex on thirty-day notice.

The number that matters is not the retainer. It is the year-one delta. Founders who run the math on the spreadsheet come out two to four hundred thousand ahead in the first twelve months, on a Series A budget, while shipping more sales touches, more content, more support coverage, and faster reporting than the comparable in-house team would have shipped. The capital saved is runway. The runway is the next product bet. That is the trade.