Decisions that Remember

Ask a senior leader what mistakes their team makes, and they'll answer without hesitating — people miss the bigger picture, they handle what's in front of them and let the systemic consequence slide. Ask a new manager in their thirties for the biggest mistake they've ever made at work, and the room goes quiet. Not from embarrassment. They genuinely can't think of one — not one that landed, that was unambiguously theirs, that taught them something no course could.

Both are describing the same organisation. Both are right. The gap between them isn't a perception problem — it's what happens when an organisation gets very good at protecting its most promising people from the exact experience the role will eventually require them to have survived.

It isn't one failure. It's four ordinary, well-intentioned structures compounding: credentials that imply readiness they don't confer, a culture that rewards escalation over commitment, KPIs that quietly narrow what counts as your problem, and the simple fact that a decision that's crystal clear in your head rarely survives being written down for a stranger to execute. None of them is a mistake on its own. Together, they produce people who reach senior roles having been systematically protected from the one thing those roles now most need them to have.

This isn't a paper about AI. But it's the reason AI is making the problem impossible to ignore. When gathering more data took real time, waiting for it was a legitimate way to defer a decision. When it takes a second, that excuse is gone. The tools got faster. Judgement didn't get a chance to.

So we built the thing that gives it one.

Last Prompt is a consequence-bearing reasoning environment — not a game, not a course, not a simulation you can pause and rewind. You get one advisor's honest, partial view of a crisis — never the whole picture, because nobody in a real organisation gets the whole picture either — you write the plan yourself, in your own words, and the world moves on the quality of your reasoning, not the confidence of your intentions. There's no reset button. What you're navigating three chapters in is the world your own decisions built.

Two things make this harder than it sounds, and they turn out to be the same problem at two different scales. First: every piece of advice you get is honest and incomplete. The specialist in the room cannot see outside their own domain — that isn't a flaw in the writing, it's the accurate shape of expertise — so the picture you're building is always partial, and synthesising it is your job, not theirs. Second: nothing resets. The decision you made two chapters ago is why this chapter is hard. Your reasoning today reshapes the conditions the next decision has to survive. That's chronosymbiosis — not a separate framework bolted onto the engine, but the same partial-information problem, running through time instead of through a room full of advisors.