Custom business software for small and mid-sized companies.

Many companies pay monthly for SaaS tools full of features nobody really uses – yet missing the ones they actually need. The way your firm is organised can only be approximated with custom fields. The reports you actually steer by live in spreadsheets synced by hand. And with every seat paid for, each additional user becomes a budget question, so half the team works from exports and screenshots instead.

Building your own used to mean months of work and budgets only large corporations could justify, so making do was the rational choice. AI coding agents have changed that calculation. With the right person steering them – someone who understands your business, writes and reviews the code, and tests what the agents produce – a tool built around the way you work can be ready in weeks, and it’s entirely yours. I’ve seen first-hand how much relief and excitement a step like this can spark at the top.

Anything where the way you work is specific enough that standard tools fight you. Some examples from my own work and from the conversations I have:

We start with how your team actually works, not with a feature list. Together we design the tool around that, and I build it – writing and reviewing the code myself, sped up by AI coding agents whose work I test. Whenever it’s possible and makes sense, we hook it into the tools you already use: your email, your files, your project system, Slack. A first usable version arrives early, and we shape the rest from real use.

What you get belongs to you: the code sits in your own repository, the data in your own database, hosted in whichever region your data needs to stay. No per-seat fees, so the whole company can work in it, not just the people with a license. And it is built on a standard, widely used stack – so any developer can continue it, not only me.

A tool like this is often what makes AI pay off in the first place. An assistant that reads your inbox and calendar, drafts follow-ups or reminds you who to call next is only as good as the system it runs on – it needs clean data about clients, deals and conversations to work from. A custom tool gives you exactly that, in a shape you control. Which is why the companies I work with usually get the foundation right first, and add AI carefully on top.

If your process is close to the standard and a good tool exists for it, use the tool – bookkeeping is the obvious example. Building pays off where the way you work is genuinely your own, where you are paying for seats and features you don’t use, or where the spreadsheets around a tool have quietly become the real system. If that isn’t your situation, I will tell you so in the first conversation.