Insight

Everyone Already Had a Spreadsheet

The argument that AI will move building out of a specialist function and into the hands of people who know why the thing is needed is not a prediction. It is a description of something that already happened, to a different competence, within living memory.

Financial modelling used to be specialist work. You queued for it. Then everyone had a spreadsheet, and within a decade the queue was gone and the modelling was being done by the people who understood the question.

That part went well. The part worth remembering is what came with it.

Critical business processes ended up living inside workbooks maintained by one person, understood by nobody else, with logic that had never been reviewed and no version anyone could point to as correct. They were discovered the way these things are always discovered: when the person left, or when the number was wrong and mattered.

The competence had been distributed. The ownership had not. And ownership is the harder half, because a spreadsheet can be handed to anyone in an afternoon while the discipline of being answerable for an outcome takes considerably longer to spread.

This is the whole of the caution about AI, and it is not a caution about the technology. An organization that distributes the ability to build without distributing the discipline of ownership does not become more adaptive. It becomes more fragmented — capability everywhere, accountability nowhere in particular, and a great deal more of it than any spreadsheet era produced.

The answer is not central control, which would give back the queue. It is settling ownership before the capability arrives, because it will arrive either way, and it arrives faster than governance can be invented for it.