The Recruiter You Keep Not Hiring

You have five open roles. You posted them six weeks ago. You have looked at maybe two hundred LinkedIn profiles in evening tabs that you keep meaning to close. You have done eleven first-round calls in the last month, eight of which you knew were a no by minute four. You have made one offer. They took the other one.

Your hiring function is one founder, a Notion doc called "Roles," and the ambient guilt of a Greenhouse pipeline you log into on Sundays. The recruiter you keep meaning to hire is not in the budget this quarter. So nothing happens, except slowly, expensively, and in the worst hours of your week.

This is the hiring function at most Series A and Series B companies. It is not a department. It is the founder, in airplane mode, doing pattern-matching on profiles at 11 PM.

The role is open, the function is not

Open your careers page. Count the roles. Five, eight, twelve. Now open your ATS and look at how those pipelines are moving. The typical pattern at a thirty-to-eighty person company looks like this. One or two roles have any real motion, because the hiring manager personally cares and is sourcing on the side. Three or four are stalled at the "we posted it and got fifty applicants, most of which are bad" stage. The rest are open in name only, waiting for a recruiter who has not been hired yet.

The work that turns an open req into a hire has a clear shape. Source the top of the funnel. Screen the inbound. Reach out to the passive ones. Schedule the calls. Run the first round. Write the debrief. Coordinate the loop. Reference the finalist. Close the offer. Nine steps, repeated across every role, every week.

At your stage, one human is doing maybe three of those nine steps for one or two of the open roles. The other six steps for the other four roles are not happening. The function is not understaffed. It is unstaffed.

Why hiring a recruiter is the slow answer

The standard fix is to hire a recruiter. In-house, loaded cost runs ninety to one hundred sixty thousand a year in the US, sixty to one hundred in Europe, forty to eighty in lower-cost markets. Agency contingent fees run twenty to thirty percent of first-year salary per hire, which on five engineering hires at one hundred eighty thousand base lands at one hundred eighty to two hundred seventy thousand for the year.

Both options have the same problem at your scale. The expensive part of hiring is not the sourcing tool subscription. It is the time between a profile appearing in a search and a calibrated, briefed call landing on the hiring manager's calendar. That gap is currently measured in weeks because nobody owns it full-time.

A recruiter hire closes the gap, eventually. The first ninety days are mostly the recruiter learning your stack, your bar, your team dynamics, and which of your hiring managers gives useful debriefs and which give vibes. The second ninety days are when the pipeline finally moves. By month seven you have a functioning hiring engine and a recruiter who is now the second-most-expensive non-engineer on the payroll. That is fine if you are hiring twenty people a year. At five to ten hires a year, the unit economics are loose.

The agency version solves the time problem and hands you the unit economics problem. Two hundred fifty thousand in fees for five hires is a real number, paid out of cash you would rather be spending on the people you hired.

What a fractional AI Hiring Department does

A fractional AI Hiring Department takes the nine-step shape above and assigns most of it to agents that run continuously, with your hiring managers and founders looping in only at the decision points.

Sourcing on its own clock. The agent runs the search every day across LinkedIn, GitHub, Wellfound, and whichever niche communities matter for your roles. Not a saved search you forget to check. A daily pass that scores new candidates against the role brief and queues the top of the list for outreach.

Calibrated outreach. Each top-of-funnel candidate gets a personalized first message in the voice of the hiring manager, referencing something specific from their profile or work. Reply rates on calibrated outbound from agents land in the high single digits to low double digits, against the one to three percent typical of bulk recruiter spam. The math works because the agent reads every profile before writing the message.

Inbound screening with a real rubric. Every inbound application gets read against the same rubric your hiring manager would use. Resume, portfolio, GitHub, writing samples. The agent writes a one-paragraph summary, a yes/no/maybe verdict, and the specific evidence behind it. Your hiring manager opens the ATS and sees a shortlist of five briefed candidates, not a queue of fifty.

Scheduling without the thread. When a candidate says yes to a call, the agent handles the calendar dance, the prep email, the interview brief to the panel, and the reminder. The first time a human touches the loop is when the call starts.

Debrief consolidation. After each round, the agent collects scorecards, reconciles disagreements, surfaces the specific evidence each interviewer cited, and drafts the next-step recommendation. Loops that used to drag for ten days because someone forgot to write up their notes close in two.

Pipeline reporting that is current. Every Monday your hiring managers get a one-screen view per role. Top of funnel, in-process, finalist, offer, close. The number is current because the agent updated it overnight, not because someone hit refresh in Greenhouse.

The unit economics, with real ranges

Compare three setups for a company trying to make six to ten hires in the next twelve months.

The founder-runs-hiring version costs nothing in payroll and a lot in everything else. Six to ten hours of founder time a week, across sourcing, screening, and loops, at the loaded rate of someone who should be selling and fundraising, lands somewhere between ninety thousand and one hundred eighty thousand a year in opportunity cost. Plus the cost of every role that stayed open three months longer than it should have.

The in-house recruiter version costs ninety to one hundred sixty thousand in loaded comp, takes ninety days to ramp, and produces real output starting in month four. Good answer if you have twenty hires a year ahead of you. Loose answer if you have six.

The agency version costs nothing until a hire closes, then takes twenty to thirty percent of base for each one. Five hires at one hundred eighty thousand base each lands at one hundred eighty to two hundred seventy thousand for the year, paid in cash. Plus the candidate quality drift that comes from recruiters who get paid only when someone signs.

The fractional AI Hiring Department version installs in a 14-day sprint and runs ongoing. Sprint cost lands in the low to mid five figures depending on scope. Monthly run cost is closer to a single senior contractor than a full recruiter. The agents handle sourcing, screening, scheduling, and debriefs across every open role at once, not one at a time. Your hiring managers spend their time in the actual interviews and the actual closes, which is the part of hiring that needed humans in the first place.

What changes after the sprint

Picture the same Monday morning, fourteen days after the sprint goes live.

The careers page has the same five roles on it. The pipelines look different. Each role has a current top-of-funnel number, ten to thirty briefed candidates in outreach, four to eight in active screen, two to three in panel loops, one in offer stage. Every name in every stage has a one-paragraph brief attached and a reason it is at that stage.

You open Slack at 9 AM and the agent has posted the overnight summary. Three new finalists surfaced for the staff engineer role, two of them currently at companies whose technical bar you trust. One candidate in the GTM lead loop went quiet after the panel and the agent has already followed up with a calibrated re-engage message. The offer for the senior PM role went out Friday evening and the candidate replied at 7 AM with two clarifying questions, which are sitting in a draft response your head of product can edit and send.

You have not opened LinkedIn search this week. You have done four interviews, all of which were worth doing, because the agent already did the screening pass that used to chew up the first eleven minutes of every call. The five open roles are moving in parallel instead of in sequence. The recruiter you kept meaning to hire is no longer on the budget line for next quarter, because the function exists without that hire.

If your hiring function is currently a founder, a Notion doc, and a slow ATS, the version where it runs continuously is fourteen days away. After that, the only part of hiring that needs you is the part you wanted to do anyway. That is what a hiring department is supposed to feel like.