The Unfilled Requisition Is Already Costing You

Every operations leader has a version of the same spreadsheet. One tab lists the roles that are budgeted. Another lists the roles that are actually filled. The gap between them is treated as a saving.

It isn't. An empty seat is not a line item you get back at the end of the quarter. It is work that still has to happen, redistributed to people who were already at capacity, at a quality level nobody signed off on.

The arithmetic nobody runs

When a mid-level ops requisition opens, the budget stops spending. That is the visible number and it is the only one most companies track. What continues, invisibly, is everything the role was supposed to own.

Invoices still arrive. Vendors still need chasing. Reconciliations still fall due. Customer escalations still land, and someone still answers them, usually the person one level up, whose actual job is judgement, not throughput. The requisition sits open for three to six months. Ramp takes another two to four. For most of a year, senior time is being consumed by work that was explicitly scoped below it.

That is the real cost, and it is not a salary saving. It is a downgrade in what your most expensive people spend their hours on, plus a slow accumulation of small operational errors that nobody has bandwidth to catch. Late supplier payments. Missed deduction claims. A compliance log that stops being current in March and gets reconstructed in September.

None of that shows up as a cost of vacancy. All of it is one.

Why the seat stays empty

The usual explanation is budget. Sometimes it is. More often it is supply.

Mid-market companies are competing for senior operational talent against employers who can pay more, promote faster, and offer a bigger platform. The candidates who can genuinely run a multi-jurisdiction reconciliation process, or own a marketplace account end to end, have options. If your operation spans several platforms, several languages, and more than one regulatory regime, which describes most Asian mid-market operators, the shortlist gets shorter still.

So the requisition stays open. Then it gets downgraded: hire someone more junior and train them up. That is a legitimate strategy, but it does not solve this quarter's problem, and it moves the supervision burden onto exactly the senior person you were trying to free up.

The third option is to leave the role unfilled indefinitely and absorb the work permanently. That is the most common outcome and the least examined one.

Reframing the decision

The question is usually posed as: can we afford this hire? A more useful question is: what is this work worth having done properly, and what is the cheapest reliable way to have it done?

Split the role honestly. Most mid-level operations positions are a bundle of two very different things.

The first is systematic work. Structured, repeatable, high-volume, rule-governed: order processing, invoice matching, account reconciliation, listing management, compliance tracking, routine reporting, documentation. It is demanding work and it has to be right, but it does not require a human to originate the decision. It requires a human to be accountable for the exceptions.

The second is judgement. Edge cases, regulatory nuance, escalations where the rules do not neatly apply, decisions with commercial consequences. This is the part you were actually paying a senior salary for, and it is the part that cannot be automated away.

Bundled together, that combination is scarce and expensive to hire. Separated, it is much more tractable.

What h.work does with the split

h.work productizes that separation. You deploy a named AI Specialist trained for the specific role, not a general-purpose assistant, but a Specialist scoped to account service coordination, or cold chain monitoring, or food safety documentation, or financial reconciliation. It works inside the channels you already use: Slack, Teams, email, WhatsApp, WeChat, LINE, your ERP, your CRM. There is no new software for your team to learn.

Behind it sits the judgement layer. Every Specialist is supervised by senior credentialed practitioners, identity-verified and credentialed through Humanity. Routine output is continuously monitored. Consequential decisions are routed to an expert for review before execution. Every action is logged and auditable. When an expert corrects something, the correction becomes training data, so the same error does not recur across the next hundred instances.

The economics land at 20-40% of the fully loaded cost of the internal hire, anchored to the tier of role you would otherwise fill. Junior-scope operations work runs $1K-$3K a month. Domain operations work with real judgement content sits at $3K-$6K. Senior individual contributor scope, regulatory reporting, complex compliance, board-facing operations, runs $6K-$12K. Monthly contracts, 30 days' notice, no setup fees, first month a paid trial at standard pricing.

Deployment is measured in hours rather than quarters. That is the part that changes the arithmetic on an open requisition: you are no longer choosing between paying a full salary and absorbing the work. You are choosing between absorbing the work and covering it at a fraction of the cost, this month.

What this is not

It is not a claim that you should stop hiring. Some roles are irreducibly human: anything that leads creative strategy, anything where the relationship is the product, anything where the judgement content is the majority of the job rather than the exception path. For those, hire, and take the time it takes.

It is also not a claim of full autonomy. AI Specialists escalate rather than improvise on high-stakes calls, and that is deliberate. Fully autonomous AI on consequential decisions is a risk transfer, not a cost saving. The expert review layer exists precisely because the alternative, unsupervised automation on work that has regulatory or commercial teeth, fails in ways that are expensive and slow to detect.

What it is: a way to stop treating an unfilled requisition as neutral.

The practical next step

Take your three longest-open operations requisitions. For each, write down what percentage of the role is systematic throughput and what percentage is genuine judgement. Most operators land somewhere between 70/30 and 85/15, and are surprised by it.

Then ask what it would cost to cover the throughput portion properly, with senior oversight attached, starting this week, and compare that to what nine months of an empty seat has already cost you in senior hours, deferred work, and errors you have not found yet.

The seat is not saving you money. It is just spending it somewhere you are not looking.