Why do organisations miss obvious risks?

Maretha Smit · 28 July 2026

A garden spade and fork resting on a lawn

Photo: Erich Ferdinand, CC BY 2.0, via Wikimedia Commons. Cropped.

In May 2026, Starbucks Korea launched a promotion called Tank Day on the anniversary of the Gwangju Uprising, which the military put down with tanks. The slogan came from an AI tool. Most Korean adults would have picked up the reference without effort. Nobody in the approval chain did.

The response was to close more than two thousand stores early and put staff through mandatory history and social sensitivity training. Those staff had nothing to do with the decision. Performative redress usually travels downward like this, and it leaves the room where the failure happened entirely unexamined.

What interests me is the diagnosis underneath it, because the company clearly decided this was a problem of attention. Somebody should have checked. And so the remedy was a checklist and a longer approval chain.

Attention and recognition are different things. Attention decides whether anyone looks. Recognition decides whether looking produces anything, and recognition is pattern-matching assembled out of the particular life a person has lived. You can raise attention with process. You cannot raise recognition that way, because process adds steps rather than range.

Ross Ashby put the constraint plainly in 1956: only variety can absorb variety. A system can regulate its environment only if it holds at least as much internal range as that environment throws at it. An organisation working in a market whose history it does not contain is operating with less range than the thing it is trying to manage.

Two things are narrowing that range at once, which is why this matters more now than it did five years ago. AI is shortening review chains, so fewer people see each piece of work before it leaves the building, and every one of them was a set of eyes. Meanwhile the people who remain tend to resemble each other, so their blind spots sit in the same places. Ten similar people reviewing a decision have the detection range of one.

Gillian Tett’s distinction between the bird’s eye view and the worm’s eye view is useful here. Senior roles train you to see the shape of the whole thing. The knowledge that would have caught Tank Day sits further down, where observation arrives slowly and in no format anyone has requested.

Which leaves two questions worth asking of your own organisation. Does the range exist anywhere inside it at all? And does anyone senior ever go and find it?

Read this way, inclusion stops being a matter of fairness and becomes a specification — the range of patterns your organisation can recognise at all. That version of the argument does not collapse the moment somebody produces a quarter where it appears not to have paid.

This is general information, not advice for your organisation.

Bring us the disruption you’re facing.

Let’s talk