At a company of 100 to 250 employees, work is divided across departments, but within those departments many tasks sit with one person. There is a customer service employee who knows the difficult cases. A controller who knows the reporting quirks of three systems by heart. An HR employee who knows exactly which contract clause belongs to which job group. That is a different situation than at a team of ten, where everyone does a bit of everything, and also different from an organization of a thousand employees, where processes have usually already been laid down because the size forces that. At 100 to 250 employees, the organization is large enough to specialize tasks, but often not yet large enough to also document those tasks. The work exists, the knowledge about it sits in one head.
That has direct consequences for what AI can take over. A task that is written down in a procedure, with fixed steps and fixed criteria, is relatively easy to hand over: the pattern has already been made explicit. A task that only exists in the head of one employee first has to be spoken out loud before a system, or a colleague, can take it over. That speaking out loud is exactly where a work scan begins: not guessing what happens, but describing the work task by task so it becomes clear what is routine and what is improvisation.
At this company size, a recurring pattern is that multiple systems have grown up side by side. A CRM for sales, a separate package for invoicing, an Excel file for staff planning that no one dares to replace. Each system works on its own, but the link between systems is missing or has been made by hand. That is a different threshold than at a small company, where usually one system does everything, and different from a large organization, which is usually already in an integration process. What that means concretely for transferability is described on the page about tasks that get stuck because systems are not connected to each other: a task can be simple in content and still remain human work, purely because someone now manually types over data from one system to another.
Within this company size there is often also a shift between departments that work with fixed products and departments that deliver customized work. A back office that processes the same five types of requests lends itself better to being taken over than a team that puts together a different solution for every customer. Exactly where that boundary lies depends on how much variation there really is in the customized work, and that is explained on the page about what changes when a large part of the work consists of customization per customer. The same applies to customer contact: a team that mainly answers repetitive questions sees a large part of that work shift toward a system with human oversight, while a team that negotiates or advises sees that shift much less. That distinction is worked out on the page about the work that changes when customer contact fills a large part of the day.
A task that at this company size often still sits with a single employee is sending welcome messages to new customers. It feels like small work, but it happens daily, almost always follows the same pattern and rarely depends on exceptions. Whether that pattern is strong enough to run without constant oversight, or whether there is reason to have a human approve it, is described on the page about sending welcome messages to new customers. It is a small example of a larger pattern: many separate tasks at this company size are each small on their own, but together add up to a substantial part of office hours.
Because companies of this size regularly introduce new systems, put together new teams or revise processes, a measurement at one moment is never the end point. A task that today still sits with one person may within a quarter have been laid down in a procedure, and with that, what AI can take over from it also shifts. What a second measurement concretely shows compared to the first is explained on the page about what a remeasurement of the work distribution yields.
The distribution of work at a company of 100 to 250 employees, with specialization per person and systems that don't always connect to each other, calls for a picture per task rather than a general estimate. The free quickscan gives a first direction: twelve questions, without an account, with an indication of what part of the hours in your profile can be taken over by AI today. The full work scan, which lays out the work task by task on eight axes, is still under construction. We therefore do not yet offer it, but we will announce here when it becomes available.
Decisions about the deployment of personnel do not fall under this scan and not under AI TO FTE. Separate legal requirements apply to that, and those are not covered here.
Vraag maar. Ik ken de kennisbank van deze site; wat ik niet weet, zeg ik erbij.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.