Most institutions have a governance problem they cannot see on an organisation chart.
The chart shows authority. It shows who approves what, where escalation goes, and which committee owns which decision. What it does not show is whether the person holding authority over a decision is the person best equipped to make it well. That second question is rarely asked, because the first one looks like an answer. The approval matrix is complete, so the governance is assumed to be sound. Completeness and soundness are not the same thing.
I call the gap between them knowledge-authority misalignment: the structural distance between who owns a decision and who understands it well enough to own it well.
Where the distance opens up
The misalignment is not random. It opens in predictable places, and almost always for the same underlying reason.
A technology risk decision is ratified by someone whose last genuinely technical role was a decade ago. A model deployment is approved by a committee selected for seniority rather than quantitative depth. A new product is signed off by a leader whose career was built in distribution, not in the operational risk that the product creates. In each case the authority is real and correctly documented. The knowledge required to exercise it is somewhere else in the building, usually several levels below the signature.
None of this involves bad intent or weak individuals. It is institutional inheritance. Authority was allocated when the organisation was smaller, or when the person in the role had also built the underlying knowledge themselves. At the time, authority and understanding sat in the same place. Then the institution grew, the technology changed, and the product became more complex. The knowledge moved. The authority structure stayed where it was. The gap is the residue of conditions that no longer hold.
Why it stays invisible
Knowledge-authority misalignment is unusually good at hiding, because nothing in the normal operation of an institution surfaces it.
A decision made by someone without the relevant depth does not announce itself. The approval is given. The form is signed. The committee minutes record a discussion. From the outside, and from the inside on a normal day, this is indistinguishable from a decision made with full understanding. The framework was followed in both cases. The documentation looks identical. The difference is in something that leaves no trace: whether the person approving actually understood what they were approving, or relied on a recommendation, and on the reasonable assumption that someone below them had done the thinking.
The misalignment becomes visible at exactly one moment, and it is the worst possible one. Something goes wrong. The decision is examined. And it becomes immediately clear that it was made confidently by someone who did not fully grasp what they were deciding. The judgment was not there. It was never there. The institution simply had no mechanism that would have shown the absence before the failure did.
What AI changes
The misalignment has always existed. What changes its significance is the speed and technical density of the decisions institutions now face.
When decisions were slower and less technical, the distance between authority and understanding was survivable. A senior approver without deep technical knowledge could rely on time, on questions, on the judgment accumulated over a long career in a stable domain. That cushion is thinning. Model-driven decisions are faster. The questions they raise are more specialised. The career judgment of a senior leader, built in a previous technical era, transfers less cleanly to a domain that did not exist when that judgment was formed.
This turns an occasional problem into a structural one. The more an institution pushes consequential decisions through automated and semi-automated systems, the more often it asks someone with formal authority to stand behind an output that they are not positioned to interrogate. The approval still happens. The understanding behind it gets thinner each time the underlying technology advances and the authority structure does not move with it. The institution accumulates decisions that are properly authorised and inadequately understood. It has no running measure of how large that pile is becoming.
Looking for it deliberately
The reason knowledge-authority misalignment persists is that no standard process is designed to find it. Reviews check whether authority was exercised. They do not check whether the authority was matched to the understanding the decision required.
Closing the gap starts with asking a question the org chart cannot answer. For a given class of consequential decision, who actually holds the knowledge to make it well, and how far is that person from the authority to make it? Sometimes the answer is reassuring. Often it is not, and the divergence is the finding. The institution discovers that for an entire category of decision, the people who understand it best have no authority over it, and the people with authority last understood it in a previous form.
That divergence cannot be closed by moving every decision down to the people with the deepest knowledge. Seniority exists for reasons that are also real: accountability, breadth, the ability to weigh a technical judgment against the wider interests of the institution. The point is not to collapse authority into expertise. It is to build the connection between them deliberately, so that the person who signs is genuinely informed by the person who knows. The channel between them has to be strong enough to carry doubt, and not just conclusions.
An institution that builds this connection is not adding process for its own sake. It is making sure that its approvals mean what they appear to mean. An approval given by someone who understood the decision, or was properly informed by someone who did, is evidence. An approval given across an unexamined knowledge gap is a signature. The two look the same on the page. Only one of them will hold when the decision is tested.
The question worth sitting with is not whether an authority structure is documented, but whether, for the decisions that matter most, authority and understanding still sit close enough together to produce judgment rather than just sign-off.
I write about governance, risk, and the decisions institutions find hardest to make. If this is relevant to a problem you are working through, reach me at aan@asifahmednoor.com.
← Back to writing