Preparedness Is a Judgement Problem Before It Is a Capability Problem

Preparedness is usually discussed in terms of plans, capabilities and resources. The more consequential question sits underneath all of them: the quality of judgement being exercised before anything goes wrong. By the time a shock lands, the decisions that matter most have already been made. Investments have been prioritised, suppliers selected, warnings interpreted, assumptions embedded. Whether an institution is prepared is largely a verdict on judgements taken months or years earlier, by people who were confident at the time.

That is worth stating plainly because poor judgement is easy to caricature as incompetence. In practice, it more often develops slowly inside capable organisations staffed by intelligent people. Assumptions that once proved reliable harden into accepted facts. Familiar sources are weighted more heavily than unfamiliar ones. Contradictory evidence is filed as an exception rather than read as a signal that the situation is changing. Successful exercises reinforce confidence in the existing approach. None of this looks like failure from the inside. It looks like experience.

The problem is compounded by how information moves. As evidence travels upwards, it is compressed and filtered. Operational uncertainty becomes a project status. Competing interpretations become a single recommendation. Nuance is reduced to a red, amber or green indicator. Some of this compression is necessary, because senior leaders cannot personally examine every piece of underlying evidence. But each layer of interpretation puts distance between the decision-maker and the reality the decision is meant to address. The result can be an internally coherent view of preparedness that is increasingly detached from external conditions. Agreement then starts to feel like evidence. When colleagues, advisers and models all point in the same direction, confidence rises, even though they may be drawing on the same information, working from the same assumptions and overlooking the same people. An institution can become more certain without becoming more correct.

The people who decide whether your plan works are usually not in the room

Most financial services preparedness plans depend on actors outside the institution that wrote them: suppliers, infrastructure providers, contractors, local authorities, employees whose families will be affected at the same time as their employer, customers who may or may not behave as modelled. Each assumption about these actors looks reasonable from the centre. Whether it holds depends on their incentives, constraints and experience, not ours. This is why consultation is not the same as evidence. The question is not whether stakeholders have been asked, but whether what they said was allowed to change the decision. Exercises frequently test execution without testing the judgement underlying the plan. Responsibilities are confirmed, escalation paths work, communications flow. The harder assumptions, whether the public will trust the message, whether a critical supplier will honour obligations to us when three other clients believe they have priority, whether local capacity actually matches what national plans expect of it, often go untouched. A successful exercise should increase confidence, but only to the extent that it created a genuine opportunity for important assumptions to be disproved. Otherwise it is organisational reassurance dressed as rigour.

AI raises confidence faster than it raises competence

AI compounds this. A fluent summary can conceal gaps in the source material. A precise recommendation can rest on uncertain assumptions. By the time a synthesis reaches a board member, the people closest to the underlying evidence are several layers removed and not in a position to flag what has been omitted. The risk is not that AI produces the wrong answer. It is that decision-makers feel more certain because the answer is presented with greater clarity and authority than the evidence warrants. AI should strengthen human judgement, not stand in for it, which requires leaders to remain actively curious about what has been excluded, compressed or assumed.

What to actually do

Five questions are worth putting on the table before any significant preparedness decision. Which assumptions would cause our plan to fail if they proved untrue, particularly about human behaviour, supplier capacity and public trust. Who outside this room could tell us we are wrong, and has their evidence been allowed to alter the decision. How many stages of selection sit between us and the original evidence. What evidence would genuinely reduce our confidence, and if none would, are we still testing our judgement or defending it. How will we recognise that reality has changed.

The greatest threat to preparedness is rarely the shock nobody anticipated. More often it is the strong conviction that we already understand how people and systems will respond. That conviction is the thing to interrogate first.

This article draws on a piece by Polar Insight CEO James Tattersfield, originally published by the National Preparedness Commission: Who gets to decide that we are prepared?