I Did It Properly. Alone.
Fifteen years of contracting teaches you a pattern.
You go in, you look around, and within a few weeks you can see it. The shortcuts taken five years ago that nobody's fixed. The manual process that should have been automated a decade back. The thing that would take a month to do properly and has instead cost the business quietly, consistently, ever since.
Email templates that only a developer could edit. No automated testing. No appetite to change either. Style over substance, slick marketing over real infrastructure, short-term thinking dressed as pragmatism. You've seen it once, you've seen it a hundred times.
And if you care — and I always cared — you try to help plug the gaps. But some gaps are too many and too deep, and the honest truth is that most organisations aren't interested in longer-term fixes. They can't change, or they won't. The compounding cost of decisions never properly made sits there, invisible to everyone except the person who just walked in and can see it clearly.
That pattern was the brief for ka-do. Not a market map. Not a user interview. Fifteen years of watching what happens when you skip the upfront thinking — and a decision, when I started building something of my own, that I wasn't going to do it that way.
The method matters here.
I used AI the way a well-resourced team uses people. Gemini for divergent thinking — bouncing ideas around, seeing what lands, exploring without commitment. Claude for pressure-testing — challenging decisions, stress-testing thinking, asking whether something is worth building before building it. Between the two: a UX designer, an architect, a senior developer, available at every stage. Not faster shortcuts. Affordable rigour.
Proper tickets. Proper UX review. Iterative development where each decision was made consciously, for the long term, not patched in because the sprint demanded it.
The same standard I'd wished my clients had applied to their own systems.
Can you build a serious SaaS product alone?
The honest answer is: it depends on what you mean by alone. If you mean without a team — yes, increasingly. The tooling has changed. AI pair programming, affordable infrastructure, no-ops deployment, self-serve legal and billing — the fixed costs that used to require a team are mostly gone. A solo founder in 2025 has access to capability that a five-person startup didn't have five years ago.
What you can't replace is thinking. The risk of building alone isn't the code — it's the echo chamber. Every decision unchallenged, every assumption unquestioned, every shortcut unnoticed. The method matters as much as the output. Using AI as a thinking partner — not just a code generator — is what makes the difference between a solo project and a solo product.
It's still hard. Properly doesn't mean perfectly.
One afternoon's UX review produced fifteen tickets. The schedule autofill engine needed a fundamental rethink three weeks before launch. The front door — the pricing page, the onboarding, the thing that converts a visitor into a user — got done last, because the interesting problems got done first. I recognise that pattern too.
But the corners I didn't cut weren't process corners. They were thinking corners.
Effort as cognitive load, not just time. The schedule as the real product, not a feature. Smart Capture as the entry point, not an afterthought. Meetings as first-class objects in a task system. None of those were obvious at the start. They emerged through use, through iteration, through sitting with the problem long enough to understand it properly.
Ka-do is what it is because of the time taken to think it through.
Not a task app with features bolted on. A product built the way I wish every system I've ever walked into had been built — with the upfront thinking done, the long-term decisions made consciously, the automation in place so the humans can do the work that actually matters.
I may have shipped sooner. But I wouldn't have shipped this.
Steve Drew is the founder of ka-do — a calm task and planning tool for busy professionals.
