The Assumption Underneath

Most of the concern about capability right now runs in one direction. Organizations worry about losing something, that experience leaves, judgment thins, and the ability to do the work quietly disappears. That is a real risk, and it is the easier one to see.

There is a second way capability fails, and it looks nothing like decline. A capability can keep getting stronger while the assumption that made it valuable goes obsolete underneath it. Investment continues, the metrics improve, and the people doing the work keep getting better. Nothing about that feels like a problem while it is happening.

When capability declines, it shows up in the work. Quality slips, errors increase, and things take longer than they should. The organization finds out because the work tells it.

The other failure gives the organization nothing to find. Output rises, efficiency improves, the team performs, and every review confirms the decision to keep going. The organization is getting measurably better at something that matters measurably less, and every improvement makes the commitment harder to question.

I have watched this happen from inside an organization, where the numbers kept saying the investment was working.

How an organization measures a capability deepens this vulnerability. Most evaluate by what the capability returns. Cost per unit, throughput, efficiency gained, output produced. Those measures answer whether the capability is performing well. They are silent on whether it should still exist. A capability can generate an excellent return right up until the thing it produces stops being needed, and the return will look strong the entire way down.

Whether a capability still earns its place is a separate question, and answering it requires knowing what has to remain true for the capability to matter. Most organizations have not asked.

The evidence is usually available somewhere, but it rarely sits where the decision gets made. Someone closer to a customer, a supplier, or a piece of the market has noticed the thing that changed. That observation travels poorly, because it contradicts a set of numbers that all look good, and because the person holding it is not usually in the room where capability investment is decided.

This matters more now than it did. Assumptions are going obsolete faster than the capabilities built on them can be rebuilt, and a capability can be undermined without ever getting worse, because the conditions that made it valuable changed around it. Organizations are also making large capability commitments right now, at speed, based on what they expect AI to change about the work. Those expectations are assumptions, and most of them have not been written down.

We call this entrenchment. A capability continues to strengthen past the point where the assumption beneath it still holds, and everything the organization can observe confirms that the investment is working.

Some of it is genuinely difficult to see. There is no incident, no missed target, no drop in performance to investigate, and the capability that will eventually cost the organization the most is, at every point along the way, one of the things it does best.

But not all of it is hidden. Sometimes the change is obvious from outside the building. A capability that once required a process, a department, or an entire company becomes something a person can now do in an afternoon. The market can see it. Customers can see it. The people doing the work can usually see it first. What happens instead is that the organization keeps reporting on how well the capability is performing, because that is the question it knows how to answer, and the harder question never gets asked by anyone with the authority to act on it.

That is not a detection problem. It is a choice, whether or not anyone experiences it as one, and the cost lands on the people whose work was built around the assumption.

Which is where the examination has to start. Not with the capabilities that are struggling, but with the ones nobody is worried about, asking what would have to remain true for them to keep being worth what they cost. Some of those answers will still hold. Others will rest on conditions that quietly stopped being true, and the organization has been getting better at the thing ever since.

The capabilities worth examining are the ones performing well.