One recruiter, twelve open reqs

Take an honest inventory of a recruiter's week and the split is grim: roughly 70% goes to mechanical throughput — reading applications, sending the same three emails, coordinating interview times, updating trackers — and maybe 30% to the work clients actually pay for: judgment, candidate relationships, and closing. Add a req and the mechanical share grows first, because every open role generates volume whether or not anyone attends to it.

That's why agency capacity has always scaled linearly with headcount, and why margins compress the moment a client adds roles. The bottleneck was never talent or effort. It's that the industry priced judgment but sold labor.

With an agent running the mechanical layer, the twelve pipelines advance simultaneously. Every applicant across every req is screened against that client's specific bar within minutes. Outreach goes out same-day. Async interviews run overnight. Scheduling handles itself. The recruiter's morning starts with twelve queues of pre-worked decisions — evidence attached — instead of twelve piles of raw applications.

The recruiter's job becomes pure operator work: review the borderline calls, prep the shortlists, run the client conversations, close the offers. One person genuinely can hold twelve reqs this way, because the part that scaled badly — the throughput — is no longer theirs. The part that's left is the part that was always the actual product.

Clients don't see your tooling. They see that every applicant to their role got a decision within a day, that the shortlist arrived Thursday instead of week three, and that each recommended candidate comes with evidence mapped to the bar they agreed on. Speed reads as competence; receipts read as rigor. Both compound into renewals.

The bar itself becomes a client artifact. Agreeing on written must-haves at kickoff — because the agent needs them — forces the alignment conversation most engagements skip, and kills the "send me more candidates, no not like that" loop before it starts. Agencies tell us this changed their client relationships more than the speed did.

None of this works if the agent is a black box the recruiter can't override. Every automated decision in Tandem is inspectable and reversible: a recruiter can pull a declined candidate back, adjust a client's bar mid-engagement, or take over any candidate conversation by hand. The agent drives the default path; the human owns the exceptions.

That division — automation for the throughput, judgment for the edge cases — is the actual product of an agent-run agency. Twelve reqs per recruiter isn't the ceiling, either. It's just where the current cohort stopped counting.