Of the seven dimensions in the maturity assessment, organization is the dimension that is most underestimated. Where IT infrastructure and data management can be pointed to as systems, connections or datasets, organization concerns something less tangible: who decides, who is responsible when an AI application does something no one foresaw, and whether there is a place where such questions come together. Many organizations start pilots without these questions being answered. That is precisely why pilots get stuck — not because the technology fails, but because there is no one who has to account for the outcome of the technology.
For this dimension, the maturity assessment does not look at what an organization says it wants, but at what has been structurally established. Is there a role that owns AI initiatives, or does that depend on whoever happens to be enthusiastic at a given moment? Is there a process to evaluate a pilot before it is scaled up, or is that improvised case by case? Does a team leader know who to turn to when an AI application makes a decision he cannot explain? These are not questions about ambition. They are questions about structure, and structure is precisely what makes the difference between one level and the next.
The five levels — baseline, foundation, activation, insight, intelligence — describe a build-up. At baseline, there is no shared picture of who decides on what; initiatives arise here and there without anyone adding them up. At foundation, there is a first form of ownership, often within a single function or department, but without the rest of the organization being aware of it. Activation means that decision-making about AI has been given a fixed place, with a process followed by more than one person. Insight adds to this that the organization learns from earlier initiatives — successes and failures are recorded and used. Intelligence is the level at which organizational structure itself moves along with what AI applications require, instead of AI having to fit into a structure that was not built for it.
What makes this dimension difficult is that the picture the CEO has of the organization rarely matches the picture held by the COO or the IT manager. One thinks there is clear ownership; the other does not know who to turn to. That is why the assessment works with a plot round: several people score separately, and the spread between their answers is made visible. That spread is often more informative than the average. An organization where everyone scores low at least knows where it stands. An organization where the scores diverge widely has a different problem: there is no shared picture of its own readiness, and that in itself is a sign that the level is lower than the highest score suggests.
Raising a level on the organization dimension does not depend on an investment in software or a new department. It depends on whether someone has been designated to bear responsibility for AI initiatives, on whether a process exists to evaluate a pilot before it proceeds further, and on whether the people working with AI applications know who they can turn to. For some organizations, this is a matter of formalizing something that already exists informally. For other organizations, this structure is entirely absent, and raising a level means that a decision must first be made about who is responsible for what — a decision that cannot be left to a tool.
This dimension does not stand alone. What the organization decides about ownership affects what is needed in the area of privacy and security, because without a designated person responsible, it remains unclear who grants permission for access to sensitive data. It also affects people and skills, because a process without people able to carry it out remains a process on paper. And it affects ethics, because questions about responsibility for an unexpected outcome are exactly the questions these two dimensions share. The order here is not a matter of preference: fundamental dimensions such as organization take precedence over dependent ones, simply because a decision-making process must exist before there is anything to decide.
This assessment says nothing about which part of the work in your organization is suitable for AI, nor about what that work currently costs you. It measures whether the organization can support AI — not what AI can take over. That is a different question, with a different instrument: the work scan from FTE TO AI calculates, per task, what portion of it can be taken over by AI, based on how the work is currently organized. These two pictures — what the organization can support and what the work can give up — should only be combined once you know where you stand on both.
The maturity assessment, including the plot round with multiple respondents, is still under development. Those who want to use the assessment as soon as it becomes available can join the waiting list. There is nothing to order yet; there is, however, a place to find out when that changes.
Vraag maar wat er moet staan voordat AI in uw organisatie kan landen.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.