The question of what AI can do in your company usually gets stuck on examples from elsewhere. This page shows how the work scan answers that question with your own data: from intake to report, and what a repetition adds later.
You provide what already exists: job descriptions, process descriptions, time records, task lists or exports from systems that already record the work. The more sources, the finer the breakdown into tasks. There is no fixed format required; the tool processes what comes in and asks for supplementation where something is missing.
The intake can be divided among multiple people. A management team does not have to do this alone: team leaders or department heads can each fill in their own part, after which the tool merges the pieces into one whole. This not only saves time, it also produces a more complete picture than when one person tries to reconstruct from the boardroom exactly what happens on the work floor.
The time this step takes is mainly in gathering existing documentation. Where that is in order, entering it is quick. Where functions have never been written out, it takes more time to first define them before they can be entered.
What you provide is broken down into separate tasks. A function is not a unit that AI either takes over or does not; a function consists of dozens of tasks, and each task is assessed separately. That is the difference between an impression and an outcome you can act on: see also the explanation of what it is.
That assessment is made based on fixed criteria that are the same for every task: the nature of the work, the degree of repetition, the dependence on context or judgement, and the data needed to carry out the task. This is machine work: no interviews, no sessions, no interpretation by an advisor who lets their own experience from other companies run through it. The outcome is determined by the dataset you provide, processed according to criteria that are transparent.
That makes the outcome explainable. For every task it can be traced back why it was assessed as promising, doubtful or unsuitable, and on the basis of which characteristics. Anyone who does not trust the score of a task can check which criterion was decisive, and adjust the input accordingly. An outcome that cannot be traced back is worthless in a boardroom; an outcome that can be traced back can be discussed and adjusted.
The tasks are merged again to the level at which you recognise them: per function, per team, per department. At every level it is stated which part of the work has been assessed as promising for AI, which part remains doubtful and which part falls outside consideration, with the calculation of what that means in FTE. That calculation is given as a range, not as a fixed number: the scan gives an estimate based on the data provided, not a guarantee of an outcome.
The report is built up in layers. At the top is the conclusion at company level, suitable for presenting without further explanation. Below that lies the breakdown per department and function, and at the bottom lie the tasks with their assessment and the criteria that led to it. Anyone in the boardroom who only needs the top layer can leave it at that; anyone who wants to dig deeper finds the underlying reasoning one layer down, right down to task level.
The report comes digitally, searchable and exportable, so that it can also be used outside the tool: in a board memo, a meeting with the works council, or as an attachment to a decision. More about the form of the conclusions can be found on the page outcomes.
The report is a dossier, not an instruction. It does not state what must happen, only what is possible and on the basis of which tasks. What you do next is a choice that falls outside the scan: follow a roadmap yourself, engage a partner for parts of it, or hand the report as a basis to a party that takes on the implementation. All these routes depart from the same dossier; none of them is mandatory.
A work scan is a snapshot at a moment in time. Systems change, tasks shift, and what was assessed as unsuitable last year may, through a new integration or a new version of a system, become promising after all. A repetition of the scan, with updated input, shows what has shifted: which tasks have moved from doubtful to promising, and which part of the previously calculated FTE potential has since been realised or is still open.
That makes repetition particularly useful after a period in which something has actually changed: a new system, a reorganisation, or the first round of an implementation carried out along one of the routes. The second scan requires less exploratory work than the first, because the basic structure of tasks and functions is already in place; what remains is updating what has changed.
Anyone who first wants to understand why tasks, and not functions, are the starting point, or wants to see how other companies have used the outcome, finds background in the knowledge base.
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.