Paying for the ramp

The offer letter is signed and the start date is set. On the company's books, the cost begins immediately: salary, benefits, equipment, a seat in the org chart. On the operational calendar, something else begins — a period in which the company pays full price for a fraction of the work.

This period has a name. It is the ramp, and for most professional roles it runs four to six months before the hire reaches full productivity. For complex mid-market roles — the ones that touch several systems, several jurisdictions, several kinds of judgement — it can run longer. The standard curve is well documented: roughly a quarter of expected output in the first month, half by the second, three quarters by the third. The salary does not follow the same curve. It is paid in full from the first payday.

The ramp is not a flaw in the hire. It is a structural feature of how humans enter organisations. Context lives in people, not documents. The new hire learns which spreadsheet is real, which client needs a call rather than an email, which exception is routine and which one is a crisis wearing a routine costume. None of this is written down, because the people who know it are busy doing it. So the knowledge transfers the slow way: by proximity, by interruption, by mistake.

What the ramp actually costs

The obvious line item is the output gap — the months of partial productivity from a fully paid role. Industry estimates put onboarding and early productivity loss at a meaningful single-digit percentage of first-year compensation, and considerably higher once the surrounding costs are counted.

The surrounding costs are the interesting part, because they fall on people who were already loaded:

· Manager time. A ramping hire consumes hours a week from the person they were hired to relieve. The relief the hire was supposed to bring is deferred by roughly the length of the ramp.

· Review time. Every piece of work the new hire produces must be checked by someone senior, because the whole point of the ramp is that the hire cannot yet tell which of their own judgements are wrong.

· Error cost. Some fraction of the ramp is paid in mistakes that made it past review — a mishandled escalation, a misfiled document, a client who learned the org chart was thinner than it looked.

· Option cost. The work that did not get done while everyone was teaching. This one never appears in a spreadsheet, which is exactly why it is so large.

For a large enterprise, these costs are absorbed by depth. There is always another senior person to lean on. For a mid-market operator, the arithmetic is harsher. The hire was usually made because one specific senior person was drowning. That person is now the teacher. The drowning continues, at a slightly higher salary bill, for another quarter or two.

The hire you make vs the hire you get

This is the quiet gap in every hiring decision. The role is scoped and budgeted as if the company is buying a fully formed capability. What the company actually buys is a promise: the capability arrives later, contingent on the hire working out, contingent on the context transferring, contingent on the senior person having enough hours left over to teach.

Recruiters price the search. Finance prices the salary. Almost nobody prices the ramp, because it does not arrive as an invoice. It arrives as six months of a team running at reduced capacity while paying full capacity's cost.

The rational response is not to stop hiring. It is to ask, of any given role, where the ramp has to live — and whether all of it needs to live inside a human's first six months.

A different place to put the ramp

An attended AI workforce model splits the question. The throughput work — the structured, repeatable, high-volume portion of the role — does not need to ramp inside a new hire. An AI Specialist trained for the function starts producing structured output on day one, inside the tools the company already uses. The judgement work — the consequential calls, the edge cases, the decisions where the rules do not neatly apply — is routed to a senior, credentialed human expert who reviews before execution.

Under this model, the ramp does not disappear. It moves. Context still has to transfer: the company's workflows, its exceptions, its definition of done. But it transfers once, into a configured system with a full audit trail, rather than slowly into a person who may leave in eighteen months and take the context with them. And the correction loop that in a traditional hire happens informally — a senior person noticing a mistake and mentioning it — happens formally: every expert correction is captured, logged, and fed back into the Specialist's future work. The system ramps the way a good hire does, except the curve is measured and the learning is retained.

The senior human's role changes shape. Instead of doing the work while teaching someone else to do it, they hold the judgement layer while the throughput runs. One experienced expert can oversee work across many companies, because their hours go to the decisions that actually need them rather than to the volume that merely needs doing.

What this is not

This is not an argument that the hire was a mistake, or that AI replaces the senior person. The judgement is the job — that is precisely why it should not be buried under six months of ramping throughput work. The argument is narrower: the part of a role that is structured and repeatable does not need to be paid for at a senior hire's ramp rate, and the part that is genuinely senior should not be diluted by it.

It is also not a claim that deployment is free. Configuration takes real effort, and the first weeks of any attended deployment involve the same honest questions a good onboarding does: what does done look like, which exceptions matter, who owns the call. The difference is that the answers accumulate in the system rather than in one person's head.

The question worth asking

Every hiring decision contains a hidden subscription: the ramp, paid monthly, in manager hours and deferred relief and quiet mistakes. It is possible to keep paying it the traditional way, and for genuinely senior, genuinely strategic roles that may be the right call.

But for the operational core of a mid-market company — the reconciliations, the escalations, the listings, the documentation, the coordination — it is worth asking the question plainly: where does the ramp live, and who is paying for it.

The companies that answer honestly tend to discover that the most expensive part of the hire was never the salary. It was the six months.