A leadership team formulates an ambition: growth, margin, faster time to customer. Everyone nods. But whoever reads that ambition as "we do more with fewer people" means something different from whoever reads it as "we let AI handle most of the routine work and keep people for the work that requires judgment". Both readings fit the same sentence on the same strategy page. The difference only becomes visible the moment someone has to decide: who gets the mandate over an AI system, who signs off on an outcome that a model has prepared, and who is responsible when it goes wrong.
Strategy and governance are therefore not a separate dimension alongside the others. It is the layer in which an ambition is translated into what may happen, by whom, and with what oversight. Without that translation, an ambition remains a sentence on paper.
AI is already taking over work today, in parts, not everywhere at the same pace. This happens in three forms that run through every layer of the organization:
1. AI performs the task entirely. 2. AI performs the task in part, with a human approving or rejecting, with reason. 3. The task remains human work, because judgment, responsibility, or context require it.
For processes and operations, this means a different division of work. For governance, it means something else: who holds the decision right when a task shifts from form 1 to form 2, and who notices when a task that was regarded as "human work" no longer is? In many leadership teams, that decision right is not documented anywhere. There is an ambition, there is a technology agenda, but there is no answer to the question of who may say that a team will from now on work with AI support instead of without it.
Companies where this already works have written out the ambition in layers: a vision, the state in which that vision becomes concrete (vision state), a target, and the state in which that target becomes concrete (target state). At each layer, it has been specified what AI work means for it and who decides on it. As a result, a middle manager can answer a question such as "may this team use an AI tool for initial assessments" without that question needing to go back to the leadership team.
Companies where this does not yet work often do have a strategy and an AI initiative, but no connection between the two. The strategy speaks of growth and efficiency in general terms; the AI initiative runs within the IT department or with a separate team. No one has determined which capability the organization needs to bring the two together: who tests whether an AI outcome is good enough, who owns the risk if it is not, and which decision on that must sit at leadership level and which must not.
That difference is not a matter of ambition level. It is a matter of readiness: has the organization already organized the decision rights, the division of roles, and the review moments, or does that still need to happen before work can actually shift.
The noticeable difference is not in the strategy presentation but in the moment after. A leadership team that has made the translation can answer a question about AI use in a team within the existing mandate. A leadership team that has not made that translation leaves the question unanswered, or decides ad hoc, after which the next similar question is up for discussion all over again. That does not cost money you see on an invoice, it costs freed-up capacity that never actually gets freed up because no one was allowed to decide that the work should be organized differently.
The work behind this dimension therefore does not lie in writing an AI policy. It lies in four layers that connect to one another: a vision that indicates why AI work is relevant to the ambition, a vision state that describes what that looks like in substance, a target that makes that concrete in capacity and outcome, and a target state that establishes which roles and decision rights are needed for it. Eight dimensions and five confidence gates test whether those layers actually connect or merely sit alongside one another.
If the subject moves toward personnel consequences, then this applies: separate legal requirements exist for that, apart from this assessment.
Strategy and governance do not stand apart from the rest. Whether a decision right works depends on how processes are set up to actually process AI outcomes, on whether the technology and capacity are in place to carry out those decisions, and on whether performance measurement shows whether a shift delivers the intended result. The question of how a leadership team learns to use the same words for the same ambition belongs here directly as well, as does the question of why a strategy sometimes remains a document instead of a decision framework.
The underlying question — which work in this company can genuinely be taken over by AI — is answered per task with the work scan from FTE TO AI.
The free readiness check consists of eight short questions, one per dimension, and provides a picture of where your organization is furthest along and where work still remains. The full ambition assessment, with the four layers and the five confidence gates, is under development.
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.