The Cost of Being “Surprised”

Organisations often describe adverse outcomes as surprises. The term is used to signal that an event was unforeseen, that available information did not indicate its likelihood, and that reasonable processes were followed. It carries an implicit defence. If something could not have been known, it could not have been prevented.

In genuinely unforeseeable cases, this may be true. In many others, it is not. What is described as surprise is frequently the result of signals that were present but not recognised as meaningful. Concerns were raised but not pursued. Patterns were observed but not connected. Assumptions were questioned but not revisited. The information existed, but the organisation did not have the capacity, or the willingness, to interpret it in time.

The distinction between unforeseeable and unacknowledged is therefore critical. It determines whether learning is possible, and whether the conditions that produced the outcome are likely to persist.

Traditional risk frameworks tend to focus on the former. They classify events, analyse root causes, and update controls. These activities are necessary, but they often reinforce the idea that the event itself was the primary problem. Less attention is given to the interpretive processes that preceded it. How were signals received? How were they discussed? Why were they discounted? What assumptions remained in place despite accumulating tension?

The connection to narrative and appetite is direct. Where narratives have begun to drift, signals that challenge them are less likely to be recognised. Where appetite is performed rather than lived, decisions that stretch boundaries are less likely to be questioned. In such environments, the conditions for surprise are already in place long before any event crystallises. Surprise is not the absence of information. It is the failure to act on what was already visible.

This failure is rarely dramatic. It is cumulative. Weak signals are treated as anecdotal. Exceptions are normalised. Confidence remains high because the formal artefacts continue to present coherence. By the time the situation demands formal recognition, the opportunity for early intervention has passed. What appears as a sudden event is often the endpoint of a gradual process that was neither invisible nor inevitable.

AI does not fundamentally change this dynamic. While it can surface patterns and highlight anomalies, it remains dependent on how those outputs are interpreted. If the surrounding system is designed to favour confirmation over challenge, AI will reinforce existing narratives rather than disrupt them. When models fail, the response is often to refine the model, not to question the framing within which it operates. The organisation becomes more precise in its analysis without becoming more responsive in its judgement.

If surprise is a failure of sensemaking rather than of prediction, then accelerating existing processes will not materially reduce it. The limitation lies in how signals are interpreted, not in how quickly outcomes are reported. Addressing this requires environments that preserve the lineage of interpretation, allowing organisations to see how understanding evolves, or fails to. Without that, signals remain disconnected, and hindsight continues to dominate learning. RiskMetrica is designed to make that lineage visible.

Within such an environment, signals are not treated as isolated observations. They are connected over time, revisited in light of new information, and interpreted in relation to evolving assumptions. Judgement is not hidden within artefacts but made explicit, allowing it to be challenged before it hardens into conclusion. This does not eliminate uncertainty, but it reduces the likelihood that uncertainty will be ignored until it becomes unavoidable.

For boards, the implication is that oversight must extend beyond the adequacy of controls to the adequacy of interpretation. When an event is described as a surprise, the most important question is not what failed, but what was seen and not understood, and why that gap persisted.

The cost of surprise is therefore not limited to the event itself. It includes the erosion of trust, the diversion of attention, and the reinforcement of defensive behaviours. When organisations repeatedly experience outcomes framed as unforeseen, confidence in governance declines, both internally and externally.

The question is not whether surprise can be eliminated, but whether the organisation is prepared to recognise when it is already in the process of being surprised.