What if the tool disappeared tomorrow?

I’ve been doing quite a lot of work lately to understand how I want to organise my different projects and decisions, and how AI and other tools can support that without creating more work for me to manage. One of the questions I keep coming back to is whether the system I’m building would still make sense if I had to change the tools tomorrow.

It’s easy to get absorbed in what a particular product can do. A new feature solves something that was awkward last month, or another tool promises to coordinate work without needing as much intervention. Those capabilities are useful, and I want to take advantage of them. But I’ve become wary of allowing the tool itself to define the way I work.

For example, if I have a reliable way to capture a useful observation, connect it to an earlier decision and decide what to do next, I’d like to preserve that behaviour even if the app that currently makes it possible changes. Otherwise I haven’t really built a dependable way of working; I’ve learnt how to operate a particular combination of products.

I think of it a bit like changing kitchens. I might have to find new places for things and learn how a different oven works, but I shouldn’t have to relearn how to cook. That’s the distinction I want to preserve between the way I work and the tools supporting it.

Practically, I want to be able to explain what should happen, where the important information belongs, and which decisions still require my judgement. Only then does it make sense to ask whether an existing tool already does enough, whether a small adaptation would help, or whether there’s genuinely something missing that needs building.

That doesn’t mean designing every process in the abstract or trying to make everything independent of technology. There would be little point in that. My aim is to avoid making the important parts of the process disappear whenever the technology changes.

The test I’ve found useful is straightforward: If this tool disappeared tomorrow, what would I lose? If the answer is mainly the implementation, that’s manageable. If I’d also lose the process, the reasoning behind decisions and the knowledge of what happens next, I’ve probably put too much of the way I work inside the tool.

I’m still testing this approach as I go, and it’s already changed how I evaluate new capabilities: not only by what they can do today, but by whether they strengthen a way of working I can continue to use tomorrow.