Kennisbank

The intelligence level: what it looks like in practice

Intelligence is the fifth level in hybridresourcing's maturity assessment, following baseline, foundation, activation and insight. It is the level at which the seven dimensions — from organization and IT infrastructure to data management and the dependent dimensions above them — no longer function separately, but are aligned with one another. That sounds abstract until you look at what does and doesn't happen in practice at this level.

How you recognize it

At the lower levels, AI runs in pilots, disconnected from one another, with varying ownership. At the intelligence level, that pattern is broken. Decisions about new applications are no longer made per project, but follow from a structure that is already in place: clear ownership, data that is findable and reliable, infrastructure that can handle expansion without every initiative running into the same limitations again. The organization doesn't have one successful application, but a way of working in which applications reinforce one another rather than getting in each other's way.

That doesn't mean everything automatically goes well. It means that when something does go wrong, the organization knows what caused it and how to adjust. The fundamental dimensions then carry the weight of the dependent dimensions, as they should: the foundation comes first, and only then the application that relies on it.

What's different from the insight level

At the insight level, the organization often already has a good view of what's happening: there is visibility into what works and what doesn't, and measurement and adjustment take place. What is still missing at the insight level is the way in which those insights are structurally processed into how the organization is set up. Intelligence is the step at which that insight is no longer a report, but part of the operation itself. Those who want to know more about that transition will find an explanation in how you move from insight to the next level.

The spread between respondents is usually smaller at this level than at the levels below it. At baseline and foundation, departments often score very differently, because some are further along than others. At the intelligence level, that spread is generally more limited, not because everyone agrees on everything, but because the structure itself leaves less room for large differences in perception. This is something the plot round of the assessment provides insight into: not just the average level, but also how uniformly that picture is experienced within the organization.

What the step to the next level requires

Intelligence is the highest level in this assessment, which raises the question of what remains to be done. The answer is that reaching a level is not an endpoint. Infrastructure ages, teams change, new applications bring new requirements for data and management. Maintaining this level requires just as much attention as reaching it, and the question shifts from 'how do we get here' to 'how do we hold on to this while the organization and the technology keep changing'. An explanation of that question can be found in how you keep building from intelligence.

For organizations that have not yet reached this level, it is useful to know that the road there does not happen in one step. The route runs through the levels below it, and the questions that arise at each level differ. Those wondering what the first steps look like will find that in how you move from baseline to the next level and in the step that follows after that, described in how you move from activation to the next level.

What this level does not say

Reaching the intelligence level says something about the extent to which an organization is set up to support AI applications. It says nothing about which tasks within that organization are suitable to be taken over by AI, and to what extent. That is a different question, requiring a different instrument.

hybridresourcing's maturity assessment is about the supporting side: are the organization, the infrastructure and the data management in a state where AI applications can function without running into fundamental limitations. That question is conditional on the other question, namely which part of the work itself is suitable for AI. For that question there is the work scan from FTE TO AI, which calculates per task what portion of the work can be taken over. An organization at the intelligence level has the structure in place to do something with the outcomes of that scan; those who want to know what a work scan concretely delivers once AI actually starts performing tasks will find that in what you can do with agents that perform work.