Wholesale runs on flows: goods, orders, invoices and information moving back and forth between suppliers and customers. A large part of office hours goes into work that repeats within that flow — entering and checking orders, reconciling stock against purchasing and sales, tracking pricing agreements, matching invoices with delivery notes, answering customer questions about delivery times and availability. Alongside that is work that repeats less: negotiating with suppliers, putting together an assortment, handling a customer with a difficult situation, deciding what happens when a delivery gets stuck.
Those two kinds of work drive a different outcome when an ambition assumes AI-work. Repetitive work with fixed rules is easier to take over than work in which someone repeatedly weighs what is right in this situation. The conditions that drive the outcome are not only the nature of the work, but also how clean and structured the underlying data is — an order line that is recorded differently across five systems is harder for a system to read than for a person who knows the context.
At some wholesale companies, work is already being taken over today: checking orders against purchasing terms, summarizing supplier email into an action, filling a stock forecast based on historical patterns. At other companies, exactly the same work still lies entirely with people. The difference rarely lies in the willingness to change. It lies in whether the data is in order, whether there is someone authorized to approve or reject a system's output with reason, and whether the organization has established who decides what when something goes wrong.
Three categories run through almost every process in wholesale. There is work that a system can take over without intervention, such as transferring an order line into the correct format. There is work in which AI makes a proposal and a person assesses that proposal — a credit note that is proposed but approved or sent back with reason by an employee. And there is work that remains human work, such as the conversation with a supplier who fails to meet a delivery time. Which task falls into which category differs per company and per system, and shifts as soon as data or process changes.
An ambition such as "our customer service will soon largely run automated" assumes that the company can categorize customer questions, that the knowledge behind an answer is structured somewhere, and that there is an agreement about when a question goes to a person. Without those three, the ambition is a sentence, not a plan. The same applies to an ambition around inventory management: that requires purchasing, sales and logistics data to align with each other, and that someone is authorized to correct an order recommendation proposed by AI.
The ambition assessment records such ambitions in four layers — vision, vision state, target, target state — and tests them against eight dimensions of readiness, with five confidence gates that indicate where the plan still rests on trust rather than on evidence. That same assessment translates the outcome back into capabilities: what must a role be able to do, and who gets the decision right when a system puts forward an outcome. For wholesale companies also considering how much of their profit goes into holding onto market share, it is worth reading how you weigh growth against margin in a strategy, because an AI ambition without that consideration quickly becomes an assumption rather than a choice.
Wholesale shares much with the logistics links before and after it. Anyone wanting to compare with a sector where the physical handling of goods sits even closer to the process will find that in which AI ambitions are at play in the transport sector, and anyone wanting to see how the same ambition translates into a production environment can read which AI ambitions are at play in manufacturing. The pattern repeats itself: the same ambition words, different readiness, different outcome.
Where this touches on personnel — which roles change as work shifts — separate legal requirements apply that fall outside this assessment. What the ambition assessment does do is make clear what capacity becomes available when certain work shifts, in hours and FTE, so that a management team knows what it is talking about before it talks about people.
The underlying question — which work in this company can genuinely be taken over by AI — is answered by FTE TO AI's work scan per task, rather than per role or per assumption. For a management team that first wants to know whether its own vision is measurable enough to test against, there is how you make a vision measurable.
Anyone wanting to know where their own organization stands can take the free readiness check: eight short questions, one per dimension, resulting in a picture of where you are furthest along and where you are least far along. The full ambition assessment, with the four layers and the five confidence gates, is under construction.
Vertel wat u wilt bereiken, dan kijken we samen wat daarvoor moet staan.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.