One of the most dangerous things I have ever said to a customer was: “It cannot happen again.”

We had suffered a major outage caused by a fault in an undersea cable. I went with our global sales team to apologise to a major bank. We had repaired the fault and contracted with an alternative supplier so that, if necessary, we could switch over immediately. I explained what we had done and gave them my assurance.

Two months later, I was back apologising to the same customer. Both cables shared a single point of failure. We had changed the supplier without establishing whether we had removed the vulnerability.

It was difficult to check. That was no consolation to the bank. Nor was the confidence with which I had delivered my first apology.

I think about that experience when I hear leaders talking about the world we now inhabit. War, political confrontation, AI, climate change. There is pressure to explain what it all means, reassure people and present a convincing plan. Leaders are supposed to know where they are taking the business.

But what can you honestly promise?

You cannot promise that a market will remain open, that a technology will develop as expected or that a conflict will stay within its present boundaries. You cannot promise that the business model which pays everyone’s salaries today will still work in five years.

Yet much of our leadership behaviour rests on the assumption that we can offer that certainty. We announce a strategy, build targets around it and reward people for delivering the plan. Changing direction can then look like failure, even when the evidence says that continuing would be foolish.

That creates a danger of its own. The organisation becomes committed to defending yesterday’s judgement.

A board approves an acquisition on the basis of particular market conditions. Those conditions change, but the deal has acquired momentum and the chief executive’s reputation is attached to it. A company invests heavily in an AI programme. Early results suggest that customers want something different, but nobody wants to challenge the promised return. A leadership team keeps pursuing growth in a market because withdrawing would mean admitting that its original assessment was wrong.

The external threat may be beyond your control. Your willingness to respond to it is not.

This is where I think the conversation about risk needs to go further. We spend considerable time asking what might happen to us. We should also ask whether we have built an organisation capable of changing its mind.

Can someone bring unwelcome evidence into a meeting without being accused of negativity? Can a senior leader stop a project they championed? Do people understand which decisions they can take themselves when circumstances change, or must everything travel upwards through a hierarchy designed for calmer times?

These are practical questions about how quickly a company can respond while it still has choices.

Consider two businesses facing the same deterioration in a market. In one, managers soften their forecasts because they know the chief executive does not want another difficult board conversation. In the other, they report what they are seeing and have the authority to adjust spending, stock and customer commitments. The second business may still suffer badly. But it has given itself time to act. The first has spent that time preserving appearances.

People do need reassurance. They have mortgages, families and careers invested in the organisation. But reassurance should come from knowing how their leaders will behave when things become difficult.

You can promise to tell them what you know and what remains uncertain. You can explain the choices being made and the consequences you are trying to avoid. You can make it clear that reporting a problem early is useful, and that concealing one to protect a target is unacceptable.

You can also distinguish between commitments that matter and plans that may need to change. Paying people, serving customers responsibly and treating colleagues fairly are commitments. A particular product launch, acquisition or growth forecast is a judgement about how to fulfil them. Leaders need to know the difference.

None of this removes the need to examine existential risks. It makes that examination more likely to produce action. A beautifully researched assessment achieves little if the leadership team cannot bear to alter the strategy it threatens.

My mistake with the bank was to promise an outcome that I had insufficient grounds to guarantee. Experience has made me much more careful about that.

In an uncertain world, your people need leaders whose judgement they can trust, including when that judgement changes.

What have you made it difficult for your organisation to tell you? And what are you still promising that you can no longer reasonably guarantee?