Build what you know

By Peter Springett · 18 May 2026

There's an adage that when writing a novel, you should write about what you know. It's controversial amongst writers, but some of the most famous books were drafted exactly that way from David Copperfield to The Devil Wears Prada. Take your pick.

The same rule applies to building apps. At a time when authoring software is a fundamental professional skill, a good place to start is with what you already know. Ask yourself: what is your specialism? What is the process that underpins it? What frustrates you about the tools you already use? How might your personal experience shape what you build, whether that's an application, a website, or another piece of software?

Then build it. Your first tool may have a niche audience. Perhaps it's something only you will use, or a handful of friends. Maybe it'll be adopted by millions of people and businesses. But probably not. Don't worry, that's not the point, yet.

Here's the thing. Once you understand how to build, once you've learned how to take a process and translate it into something useful, you can start talking to other experts. They are usually closer than you think: colleagues, former clients, people in your network who've complained about the same broken workflows for years. People who don't share your enthusiasm for AI, but who have their own goals, their own processes. The tools they use are limiting or too general.

Maybe they don't exist at all. Offer to build what they need. Draw out their knowledge. Understand the process they know inside out. The real lesson isn't just to start with what you know. It's that doing so teaches you how to build. Not technically, but conceptually. You learn to translate human expertise into working software. And once you can do that with your own knowledge, you can do it with anyone's. Expertise. Process. Tool.

Get that pattern right, and it doesn't matter whose expertise you start with next.