The Era of Good Judgement

On the night of 26 September 1983, Stanislav Petrov was the duty officer at a Soviet nuclear early-warning facility outside Moscow when the alarm went off. The system was unambiguous: five missiles had been launched from the United States. Protocol was clear and Petrov’s job was to report it up the chain and let the process take over.

But he didn't. Something didn't feel right. A genuine American first strike, he reasoned, would not involve five missiles. It would involve hundreds. The data was saying one thing, but his judgement and experience were saying another. He reported a system malfunction and he was right. The satellites had been fooled by sunlight reflecting off clouds. Petrov's decision almost certainly prevented a Soviet retaliatory launch and what might have come after doesn't bear thinking about.

We are living through an extraordinary moment in the history of decision-making. Artificial intelligence can now ingest more information in a second than a team of analysts could process in a year. It can identify patterns invisible to the human eye, run complex scenario models simultaneously, and produce text and visual outputs of genuine sophistication.

Judgement is not the same as analysis. Analysis tells you what the numbers show but judgement tells you which numbers matter, and when to ignore all of them. It draws on experience, years of watching how situations evolve, how people behave under pressure, how organisations really function beneath the surface. It draws on psychology, understanding that the other side of the table isn't always acting rationally, or that the client who says they want growth but in reality fears losing control. It draws on instinct and that uneasy feeling in the room that experienced advisers learn, over time, to trust.

History is full of catastrophic decisions made by people who had all the data and none of the wisdom. In 2008, every major financial institution was working from the same risk models. The models were sophisticated, but the judgement was absent. The people who saw what was coming – as documented in Michael Lewis’ The Big Short - weren't necessarily the best quants. They were often the ones who had been around long enough to know that the music always stops.

None of this is an argument against technology, which we should and will embrace. I used an AI to help me find the Petrov story. But it is an argument for knowing what that technology is for.

AI is extraordinary at processing the known. But the moments that define organisations, the crisis that comes from nowhere, the opportunity with a narrow window, the decision that looks clean on paper and feels wrong in the room, these are precisely the moments where human experience becomes irreplaceable.

What leaders often tell us, after coming through a critical moment, is that they had all the information but they lacked someone who could sit alongside them and say: here is what this means for you, given who you are, what your context is, and what you're trying to build. That requires a different kind of intelligence, not trained on data, but on decades of judgements.

The best decisions we've seen leaders make weren't made because they had more information than everyone else. They were made because, at that critical moment, they had the right people around them, with the right experience and judgement, to help them make the right call.