An AI tool is purchased, implemented, rolled out. There is budget, there is a supplier, there is a start date. What is missing: a change in who makes which decision, how information moves between departments, or where responsibility lies when the system's outcome differs from what an employee expects. The organization remains exactly as it was, with a layer of technology on top.
The assumption is that the technology does the work and the structure can remain unchanged. That holds true for a calculator. It does not hold true for a system that needs data from three departments that already communicate poorly with each other, or that proposes a decision to someone who has no mandate to make that decision.
Purchasing technology is a concrete, well-defined action. There is a quote, an implementation date, a result that becomes visible on a dashboard. Changing structure is vague: who is going to do that, when will it be finished, what is the result. Between a clear action and an unclear action, an organization almost always chooses the clear one.
On top of that, suppliers of AI systems sell their product as something that fits the existing organization. That is also their business model: the less the customer has to change, the easier the sale. The message that the organization first needs to do something itself does not fit well into that sales conversation.
And there is a third reason, perhaps the most important one: changing structure touches on power, on who decides now and who will decide later. Purchasing technology does not touch on that. It is easier to buy than to reconsider who carries which responsibility.
A number of signals that occur together more often than in isolation:
The technology is in place, but no one has been given the authority to act differently based on the outcome than before. The system advises, the person decides as they always did, and the advice disappears into a drawer.
The data the system needs is spread across departments that have no joint process for sharing that data. The system works with whatever it happens to receive, and the quality of the outcome depends on which department happened to supply it that week.
There is a pilot team that is enthusiastic, but the rest of the organization carries on as usual. What exactly is missing there can be read in the description of a pilot that gets stuck with two enthusiasts and no one else.
The pilot started without defining what success means and who assesses that. After a few months there is a result, but no one knows whether that result is good enough to proceed. That pattern is described in a pilot that runs without criteria agreed in advance.
A decision is made to roll out more broadly while the first pilot is still running on exceptions and manual corrections. What goes wrong then is described in scaling up before the foundation is in place.
And the technology works well within a single team, but as soon as the outcome is needed by another team to move forward, the process stalls. That is the pattern of a pilot that only works within the boundaries of its own team.
If two or more of these signals occur, there is a good chance the technology is taking over a task the organization has not been prepared to take over.
The hybridresourcing.com maturity assessment looks at seven dimensions, three of which are fundamental: organization, IT infrastructure and data management. These three take precedence over the others, not because they are considered more important, but because an organization without functioning data flows and without clear responsibilities has no foundation on which anything from the other dimensions can rest. Five levels, from baseline to intelligence, indicate how far an organization has progressed on each dimension.
What makes the assessment distinctive is the plotting round: multiple people within the same organization score separately, without influencing each other. The spread that results from this is often more informative than the average. If the CEO places the organization at insight level and the IT manager at baseline, that difference is itself a signal, and precisely the kind of signal that is independent of which technology is being purchased.
What a CEO specifically focuses on in this assessment, and why that differs from what a COO considers important, is set out on the pages what a CEO focuses on in AI maturity and what a COO focuses on in AI maturity.
This assessment addresses the question of whether the organization can support AI, not which work AI could take over. That is a different question, with a different instrument. Once it is clear where the organization stands on the seven dimensions, and which fundamental steps still need to be taken, the question of which part of the work is actually transferable to AI becomes relevant. That question is answered by the FTE TO AI work scan, which calculates per task which part can be taken over. The order is not coincidental: first the capacity to support it, then the takeover itself.
The hybridresourcing.com maturity assessment is under construction. Those who want to use the plotting round as soon as it becomes available can sign up for the waiting list.
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.