A staffing plan is built on assumptions about what work costs, who does it and how many people are needed for it. Those assumptions usually don't appear in an HR document. They sit in the budget, in the staffing structure, in the job architecture. When AI takes over or partly takes over a task somewhere in the company, one of those assumptions shifts without there being a moment when someone says: this no longer holds. The HR director is often not the one who causes the shift, but is the one who first sees the consequences reflected in the staffing structure.
The question the HR director asks is not whether AI is taking over work. That is already happening, task by task, at a pace that differs by department. The question is which roles are built up from tasks that are now largely or partly done by AI, and which roles are not — and whether the job architecture still shows that difference.
Work breaks down into parts: tasks AI can take over, tasks where a human approves or rejects with reason, and tasks that remain human work. A role is almost never one category. An advisor, a planner, an intake officer has a mix of the three, and that mix changes per team, per system, per amount of oversight the company wants to build in. Anyone who reasons at the role level — "this role will disappear," "that role will remain" — misses that mix and draws conclusions that don't match what happens in practice.
An answer the HR director does not accept is an estimate without a source: a percentage of roles said to disappear, presented as fact. Every figure about AI and work depends on the tasks that were examined, the pace of implementation and the level of oversight the company chooses. Without that context, a number is an assumption someone forgot to mention.
Nor does the HR director accept that an AI outcome is used as the basis for a dismissal decision. That decision is subject to its own legal requirements, and those are not replaced by a task analysis. What a task analysis does offer: insight into how much capacity is freed up in a team and where that capacity can be redeployed. What an employer does with that is up to the employer.
The financial side of the executive team often calculates in terms of fte reduction: fewer heads, lower payroll costs. The HR director calculates in freed-up hours and the question of what happens with them — retraining, redistribution, new tasks that do remain human work. That is not a difference in ambition, it is a difference in what someone measures. Anyone who looks only at the cost side misses that freed-up capacity is also an opportunity to do work that previously got no time.
The operational side of the executive team often wants to implement quickly wherever possible. HR wants to know whether the oversight that belongs to a task has actually been set up — who approves, on what grounds, and what happens if that person disagrees. A task that "AI can take over" becomes, without that oversight, a task nobody checks anymore, and that touches directly on what HR is responsible for: how work is organised in the company and who is accountable for it.
In some companies the job architecture has already been updated: tasks have been reclassified according to what AI does, what requires oversight and what remains human work, and the staffing structure has been adjusted accordingly. In other companies the job architecture looks exactly as it did three years ago, while the underlying tasks have already shifted. The difference usually lies not in ambition but in rhythm: companies that periodically test the assumptions behind their staffing structure see the shift early. Companies that don't only see it once the number of requests for redeployment or retraining suddenly rises.
The same question plays out elsewhere at the executive table, with a different weight: how a private equity partner looks at the same shift in a portfolio company, how a banker reassesses a company's cost structure when tasks fall away, and how the supervisory board fills its role in decisions about AI and work. The HR director who discusses their staffing question only internally misses that these other roles are already tracking the same shift from their own interest.
The underlying question — which work in this company can genuinely be taken over by AI, which work requires oversight and which work remains human work — is mapped task by task with the work scan from FTE TO AI, so the staffing discussion doesn't have to rest on an estimate. That is not a replacement for HR policy, it is the factual basis on which that policy can re-test its own assumptions.
A staffing plan built on old assumptions doesn't fail because it was written incorrectly. It fails because nobody knows anymore when the assumption was last checked. A free assumption check offers a first step for this: a short round in which you name your key assumptions about work and capacity, and see for each one when it was last confirmed. The full proof print, which places this alongside external sector data and the assumptions of the rest of the executive team, is under construction.
Stel uw vraag. Vaak zit de echte vraag een laag dieper — daar mag ik naar vragen.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.