The Ramp You Paid For Twice

Every hiring plan has a start date. Almost none of them have a productive date.

That gap is the most expensive line item nobody puts in the budget. A mid-market operations lead signs an offer in March, starts in May, and is genuinely useful — owning decisions, catching things before they escalate, working without a map — somewhere around September. The salary clock starts on day one. The output clock starts four months later. Both are paid in full.

Finance sees one number. Operations lives with two.

What the ramp actually consists of

Onboarding decks are the smallest part of it. The real ramp is the accumulation of context that nobody wrote down:

• Which of the eleven reports leadership actually reads, and which one is load-bearing for the monthly close.
• Which supplier will accept a late PO revision by email and which one requires it in their portal or it never happened.
• Which customer's escalations are real and which are routine noise from one loud contact.
• Where the exception queue lives, who works it, and what happens when that person is on leave.
• Which system is the source of truth when the ERP and the spreadsheet disagree — and they will disagree.

None of this is in a document. It is held by three or four people who are already at capacity, which is why the requisition existed in the first place. So the ramp is not just slow; it is slow and it consumes senior capacity from the people least able to spare it. A new hire's first quarter has a shadow cost measured in the hours their colleagues spend answering questions instead of doing their own work.

Then the second payment comes due. Attrition inside the first year means the ramp is repaid from zero. Not the salary — the context. Everything the departing hire absorbed leaves with them, and the next person starts at the same place, asking the same three or four people the same questions.

The part that never needed a person

Look closely at what fills a senior operator's first six months and a pattern shows up. Most of it is structured, repeatable and volume-driven: invoice matching, order exceptions, reconciliation, chasing acknowledgements, assembling reporting packs, tracking documentation status, keeping the audit trail current.

It is not trivial work. Getting it wrong has consequences. But it is rule-shaped work — the kind where the difficulty is throughput and consistency, not judgement. It is also, notably, the work that gets deprioritised first when the same person is being pulled into judgement calls, which is how backlogs form in teams that are technically fully staffed.

The judgement work is different in kind. Whether to accept a deduction or dispute it. Whether an anomaly is a data error or a real exposure. Whether a supplier's corrective action is adequate or cosmetic. Whether a filing is defensible. That work needs experience, accountability and someone who will own the outcome.

Conflating the two is what makes hiring feel like the only answer. You need judgement, so you hire a senior person, and then that senior person spends most of their week on throughput.

What changes when throughput starts on day two

An AI Specialist inverts the sequence. The structured work — the invoice queue, the exception list, the reporting cadence, the documentation trail — starts moving within days of deployment, inside the channels the team already uses. No new software for anyone to learn, no six-month absorption period, because the Specialist is trained for the role rather than learning the company from scratch.

The judgement layer is where this stops being an automation story. Every consequential decision routes to a credentialed senior expert before execution, and routine output is monitored continuously. The expert corrections are not just quality control; they become the ground truth that improves the Specialist's output over time. So the context that used to walk out of the building at resignation accumulates in the deployment instead.

That is the practical difference. A hire's context is personal and perishable. A supervised Specialist's context is institutional. You pay for it once.

The commercial shape follows from the same logic. Pricing anchored at 20–40% of a fully loaded internal hire, monthly contracts, 30 days' notice, no setup fee. Not because the work is worth less, but because the ramp — the four months you were going to pay for anyway — isn't in the price.

The question worth asking before the next requisition

The instinct in an over-capacity ops team is to ask who we need. The more useful question is what fraction of that role is judgement.

If the honest answer is twenty or thirty per cent — and for most mid-market operations roles it is — then the requisition is really two jobs wearing one job title. One of them needs a senior human who will own the call. The other needs reliable throughput starting this week, with someone qualified checking the work.

Splitting them doesn't reduce the standard. It removes the ramp from the part of the job that never needed one.