A strategy in healthcare often rests on an assumption about where the time of care providers and support staff goes. A large part of that time is not spent at the bedside or at the desk, but in documentation: record-keeping, reporting, handover, billing, planning and the administrative steps around an indication or referral. In addition, there is work that revolves around judgment under uncertainty: weighing a diagnosis, adjusting a treatment plan, having a conversation with a patient or family. These two kinds of work run through each other, but they do not respond the same way to automation.
What makes healthcare different from other sectors is that the outcome of much of the work is not only driven by speed or cost, but by liability, protocol and the fact that a mistake is not always reversible. That does not mean AI does not take over work here. It means that the place where this happens, and the degree of oversight that remains in place, lies somewhere different than in a sector where a mistake only costs money.
Within healthcare organizations, work can be seen to fall into three categories. Part of the administrative work — summarizing a record, filling in a template, recognizing a pattern in routine data — can be taken over by AI, with little discussion about the outcome. A second layer works with human oversight: AI makes a proposal, a doctor or nurse approves or rejects it, with reason. Think of a draft letter, a risk assessment or an initial triage. The third layer remains human work: the conversation with the patient, the responsibility for the final decision, the situations in which context and experience cannot be replaced by a pattern from data.
These three layers exist alongside each other in almost every healthcare institution. The difference between organizations lies not in whether they use AI, but in how sharply they know which part of their work falls into which layer, and whether that division still matches what is technically possible today.
The speed with which AI takes over work differs strongly between organizations, and that difference can be traced. Where records are structured, protocols unambiguous and management centralized, work in the first and second layer shifts faster: there is less room for exceptions and less resistance to a system that takes over part of the work. Where record-keeping is fragmented across systems, where teams work autonomously and where the organization serves multiple specialties with their own rules, that shift moves more slowly, not because the technology is lacking but because the preconditions do not align.
Regulation plays a separate role in this. What falls under oversight today can be reclassified tomorrow through a change in standards or certification. An assumption that currently states "this remains human work because of regulation" can lapse without notice. That is not a reason to wait, but a reason to know which assumptions hang on that clock.
A strategic plan in healthcare often contains an assumption about the required FTE capacity for administration, planning or first-line assessment. If AI can handle part of those tasks, it is not the plan itself that changes, but the assumption on which it rests. That is a different conversation than a conversation about staff: what an institution does with freed-up capacity falls under its own legal requirements regarding terms of employment and works council involvement, and that is not what this page is about. What matters here is whether the assumption underlying the strategy still holds, and who checks that.
The same shift plays out in other sectors, each with its own rhythm and its own bottlenecks: in wholesale, the strategic pressure changes mainly around inventory and order-related work, in manufacturing, the shift lies closer to planning and quality control, and in professional services, it mainly affects advisory and reporting work. Healthcare shares with those sectors that the assumption underlying a plan can lapse without anyone noticing, something a board can ask itself using the question of how you test whether your strategy still holds.
The question of which work in a specific healthcare organization can genuinely be taken over by AI, which part remains under oversight and which part remains human work, cannot be answered with a general description: it differs per department, per system and per protocol. The work scan by FTE TO AI answers that question task by task, based on the work as it is actually carried out in that organization. Anyone wondering more broadly how a strategy relates to an environment that changes faster than the plan itself will find a starting point in what to do if the market changes faster than your strategy.
A board that suspects part of its strategy rests on a lapsed assumption can make that suspicion concrete. The free assumption check is a short round in which you name your key assumptions and see, for each one, when it was last confirmed and by whom. It is not advice and not a judgment on staff, but an initial picture of where your plan still stands on solid ground and where it does not. The full strategic stress test, with sector overview, board plot and ongoing assumption list, is under development.
Stel uw vraag. Vaak zit de echte vraag een laag dieper — daar mag ik naar vragen.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.