A wholesale business runs on work that is repetitive and comparable: entering orders, matching stock to demand, tracking purchase prices, informing suppliers and customers about status and delivery times, checking invoices and credit notes, handling returns, keeping price lists and assortments up to date. Much of this work is administrative in nature but commercial in weight: an incorrect stock estimate or a too-slow price adjustment directly affects margin. In addition, there is work centred on relationships and negotiation: contact with suppliers about terms, contact with customers about customisation, assessment of new parties in the chain. These two types of work — the repetitive and the relational — together determine where the strategic pressure in this sector arises once AI takes over work.
Within wholesale trade, three categories of work run alongside one another. Part of the work can be taken over by AI: matching orders to stock, flagging price deviations, drafting standard communication to customers and suppliers. Another part shifts to work with human oversight: a system proposes a purchasing proposal or a price change, and an employee approves or rejects it, with reason. The third part remains human work, especially where it concerns negotiation, the assessment of a new supplier without a track record, or exceptions that do not fit a pattern. Which part of the work falls into which category differs per company and depends on how standardised the assortment is, how predictable demand is, and how well the underlying data are in order. A wholesale business with a narrow, stable assortment sees a different shift pattern than a wholesale business with a broad, variable offering and many customised agreements.
The shift does not proceed evenly. At one company, order processing has already largely been taken over and the remaining capacity lies in exception handling and customised customer contact. At another company, the same work is still done entirely by hand, not because it could not be done otherwise, but because the systems do not connect with one another, the data are not clean enough, or no one has been given the authority to adjust the process. That difference rarely lies in the technology itself. It lies in who within the company has permission to redistribute a task, and in how recently someone has checked whether the assumption behind the process still holds. A strategy based on a fixed amount of order-processing capacity remains standing on paper, even if that capacity has in practice already been partly freed up or is in fact still fully occupied — and it is precisely that gap between paper and practice where the pressure arises.
The core of the shift is not that a project is under way to introduce AI. The core is that an assumption underlying the strategy becomes invalid without anyone withdrawing it. A management team that relies on a fixed number of FTEs for order processing, on a fixed turnaround time for price adjustments, or on a fixed capacity for customer contact, does not automatically notice when that assumption no longer holds. The figures in the annual plan remain in place, the meeting proceeds on the basis of the old picture, and meanwhile the underlying reality has already shifted. That happens without announcement, because no one has the task of explicitly asking: does this assumption still hold, and when was that last checked.
The question of which work AI takes over is not a personnel question and not an argument for a dismissal decision. What an employer does with freed-up capacity — redistributing it, providing training, having it do something else — falls under its own statutory requirements and its own judgement. What is at issue here is more limited and more factual: which work can technically be taken over, which work requires oversight, and which work remains human work. Those facts are needed to keep a strategy up to date, separate from what management subsequently decides to do with that capacity.
This shift does not occur only in wholesale trade. Comparable patterns — part taken over, part under oversight, part remaining human work — can be recognised in how the manufacturing industry deals with comparable shifts, in how transport companies see their planning and logistics shifting, and in how professional service providers see their advisory work splitting into takeover, oversight and human work. Anyone who wants to understand exactly what a strategic assumption is and how to see whether it still holds will find that set out on the page about what a strategic assumption is and how to check whether it is still valid. And where the shift itself is already visible in disagreement within management about where the priority should lie, that is described separately on the page about what to do if management disagrees about priorities.
The underlying question — which work in this specific company can genuinely be taken over by AI, which work requires oversight and which work remains human work — is answered per task with the work scan from FTE TO AI.
The full strategic pressure test, with sector data, management's self-plot and the assumptions list side by side, is under construction. In anticipation of that, there is the free assumptions check: a short round in which you identify the main assumptions underlying your strategy and see, per assumption, when it was last confirmed.
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.