A process is more than a task. It is a sequence of steps, with handover moments, waiting times and dependencies between people and systems. If one task in that process shifts to AI, the rest of the process changes along with it, or it doesn't — and that difference is precisely what this dimension measures.
An invoice check that AI can carry out only produces freed-up capacity if the steps before and after it are set up for that: delivery in the right form, a recorded rejection criterion, a follow-up step that doesn't manually check again what has already been checked. If that's missing, the task does get taken over, but the process remains just as slow as before. The capacity is freed up, but not deployable.
In an organization where this is in order, a shifted task is immediately visible in the lead time of the process it's part of. There is one place that records which step AI carries out, which step is under human oversight, and which step remains human work — and that breakdown is used to actually plan work differently.
In an organization where this is not in order, the shift stays stuck within the task itself. The invoice check is faster, but the rest of the procurement process runs the same route as always, with the same handovers and the same waiting time. Freed-up hours from one person dissolve into the waiting time of the next step, and no one notices that anything has changed about the FTE capacity of the department as a whole.
The difference usually doesn't lie in the AI, but in how the process was already documented beforehand. Companies with an explicit process description — steps, handovers, responsibilities — can relatively quickly point out which step qualifies for takeover, which step needs human oversight, and which step remains human work. Companies where the process only exists in employees' heads have to do that work first before there's anything to shift.
On top of that, some processes inherently offer more room than others. A process with a lot of structured input and a clear right-or-wrong criterion lends itself more easily to takeover than a process that relies on explanation, negotiation, or context that isn't documented. That's not a matter of ambition, but of how the work is already structured.
Readiness of processes and operations is not about purchasing a system. The work lies in three layers:
These layers also touch on other readiness questions. Whether a takeover holds up depends partly on what the systems and existing capacity can actually handle, and whether the results are measurable depends on how performance measurement is set up for a process partly carried out by AI. Where a process runs beyond the organization's own boundaries, how partnerships are set up for a shift in work also counts toward readiness.
When a task within a process moves to AI, more changes than just who performs it. The lead time of that piece of the process changes, the margin of error changes — sometimes downward, sometimes different in nature — and the place where a human still needs to intervene shifts from executing to assessing. That's a different skill than the task itself: assessing whether an AI outcome is correct requires different attention than producing the outcome itself.
What doesn't automatically change is the department's FTE capacity. That change only occurs if the process is explicitly set up for it: fewer handovers, less waiting time, a different planning of who does which step. Without that step, the freed-up time remains theoretical. Which course an organization then takes with that freed-up capacity — a decision about workforce deployment — is up to the organization itself and falls under its own legal requirements; this dimension only describes what the process can handle, not what an employer does with it.
Whether a shift actually takes hold in practice can also be recognized by early signals that a change in the process is actually being adopted, and by how the conversation about it is conducted within the organization, for example via a way to involve the workforce without setting up a lengthy process.
Within the ambition test, process readiness is one of the eight dimensions that together determine whether an ambition that relies on AI work is also achievable. The question of which work in this specific company can truly be taken over by AI, per task and not in generalities, is answered with the work scan from FTE TO AI.
An indication of where your organization stands on this dimension, and on the other seven, is available with the free readiness check: eight short questions, one per dimension, giving you a picture of where you are furthest along and where you are least far along. The full ambition test, 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.