One Concept, One Example, One Question

The concept

Some decisions fail later because the choice was wrong, however many fail because nobody can reconstruct what the choice was based on.

The meeting moved on. The ticket was closed. The document was updated. A few weeks later, someone asks why the team chose that route, and the answer depends on whatever people can still remember about the conversation. The useful detail is usually the first thing to disappear.

A decision receipt is not a full business case. It is the small bundle of context that lets a later reader understand the decision without replaying the whole conversation.

The basic test is simple: can someone understand the answer and why it was chosen? If there is uncertainty left over, the receipt should say where it went. Follow-up work needs an owner, and the decision needs a clear reason to be looked at again.

This is useful in ordinary human work, and it becomes more important when AI is involved. If an AI tool recommends which cases meet a risk threshold, the organisation still needs a record of what the human accepted and why. Otherwise the work can look faster while making future challenge harder: the recommendation is visible, but the judgement around it is not.

The point is not to add ceremony to every tiny choice. A receipt belongs where the consequence is material, especially if someone may be affected or may need to challenge what happened. The person who can stand behind the decision should be the person who makes sure the receipt exists.

The example

Imagine a team choosing how to handle a backlog of older cases before moving to a new system.

The team could clean every record before migration, or it could move the records first and fix problems later. It chooses a middle route: clean the records that meet an agreed risk threshold, because the deadline is fixed and full cleanup would delay the move without improving the cases that matter most.

If the decision is recorded only as "use risk-based cleanup", it will not help much later. Someone reviewing the work may not know whether the team made a deliberate trade-off or simply let data quality slip under deadline pressure.

A better receipt might say:

We will clean only records that meet the agreed risk threshold before migration. The deadline is fixed, and full cleanup would delay the move without improving the cases most likely to cause harm. Lower-risk issues will be logged for post-migration review. Sarah owns the pre-migration exception list. Review this decision if the deadline changes or the risk evidence no longer holds.

That paragraph does not make the decision perfect, but it gives the team something to inspect later.

Someone can now challenge the threshold and test the report behind it. The receipt also protects the team from a common failure: treating a time-bound trade-off as if it were a permanent principle.

The obvious objection is that deadline pressure is exactly when this discipline gets skipped. That is why the receipt has to stay small. Three or four clear sentences written at the point of decision are usually cheaper than trying to reconstruct the reasoning three months later.

The question

Ask yourself:

If someone reviews this decision in three months, what will they need to know that is currently only in people's heads?

You do not need to write an essay for every choice. For decisions with real consequence, leave enough context for a later reader to understand the judgement without having to find everyone who was there. Future you will not remember as much as present you thinks they will.

Elsewhere this week…

• I published Talk me through what a good digital teammate should and should not do, which looks at how an AI collaborator can help with the work around thinking without taking over the judgement itself.

• The SecondBrain Framework update adds the Practical Review Tools index, the standing generative-AI-use preface for new writing, and recency weighting for future voice-pack updates.

Support

Everything I publish here is free to read, and always will be.

I maintain this newsletter and my website in my own time. If something here has been useful to you, and you would like to help cover the costs behind that work, you can support me through Ko-fi.

Generative AI is a powerful tool. I openly use it to conduct research, challenge my thinking, generate drafts, and review my work. AI never owns what I think, or my personal judgement. Everything I publish is a reflection of my own thoughts and opinions.