The Cost to Play Has Never Been Lower
AI has dramatically lowered the barrier to creating customised software solutions, allowing small businesses to quickly spin up tools that fit their specific needs.
For many small business owners, the ability to create customised software solutions was once a daunting and expensive endeavour. It usually involved eye-watering quotes, long timelines and the vague promise that the end result would almost match what you actually needed.
Today, AI has dramatically lowered that barrier. Small businesses can spin up tools that fit their specific needs far more quickly, and with much less commitment. The cost to play, and just as importantly the cost to fail, is lower than it has ever been. If something doesn’t work, you haven’t sunk half a year and a small fortune into finding out.
Case Study: Building PrintCommand for a 3D Printing Business
Let’s look at a practical example.
As a software engineer running a 3D printing business, I built a tool called PrintCommand to solve a very real operational problem. PrintCommand monitors revenue but, more importantly, it dynamically calculates pricing and margins in real time.
It automatically slices print files, extracts data from 3MFs and factors in variables such as filament costs (sourced from recent AliExpress orders) and electricity prices (I collect this information via MQTT from an Octopus Home Mini). As those inputs change, pricing and margins change with them. The result is far more confidence in profitability, without living in spreadsheets or repeatedly second-guessing numbers.
It isn’t the most complicated solution ever built. In fact, I imagine most engineers reading this would scoff at the idea of it taking any time at all. The point is that AI tooling has meant I can achieve this in just a few evenings and, hand on heart, eliminate a major manual step in running my business.
While this example is rooted in 3D printing, the idea behind it is hardly unique. Small businesses tend to have very specific problems, and off-the-shelf software usually gets them about 80 percent of the way there, then charges extra for the remaining 20 percent it still can’t quite do.
My Perspective
It’s worth being clear about something. This isn’t about pretending that anyone can suddenly build production-ready software with no experience (and I'd encourage anyone to ignore the vast majority of LinkedIn hype these days suggesting otherwise). AI doesn’t replace engineering fundamentals, and it doesn’t magically turn every business owner into a developer.
What it does change is the cost of experimentation. You can now try something out, get it wrong, and iterate at a fraction of the time and cost it would have taken before. And honestly, you’d arrive in front of a development company far more informed than if you’d had them scope the work for you in many cases.
Take my PrintCommand tool for example. As a software engineer, I fully appreciate that this isn’t some gold-plated, production-tier solution. But it solves a real business problem quickly. The time to get it up and running was measured in evenings, not months. If I’d built it the traditional way-or hired another engineer to do it, it would have cost me weeks of time and thousands of pounds. That’s the difference.
Yes, there’s a risk of more half-finished tools out there, but the upside is huge. We can move faster and deliver value without bending our businesses around software that doesn’t quite fit.
MVPs Really Don’t Have to Be Fragile
One of the more interesting shifts is how this changes the idea of an MVP.
Historically, “minimum viable” often meant cutting corners. Quick, brittle solutions designed to prove a point before being rewritten later, usually by someone who quietly resented the original version. With AI-assisted development, that trade-off is far less severe.
Given the right engineering background and well-structured prompts, the upfront effort required to build something solid is often not meaningfully higher than building something fragile. You can arrive at an MVP that is fast to build and structurally sound, something that can evolve rather than be immediately scheduled for a rewrite.
Summing It Up
AI is fundamentally changing how small businesses approach internal tooling. By lowering the cost to play and the cost to fail, it enables faster progress, quicker learning and better value creation.
I'm really keen to hear if other engineers are capitlising on this wave, are you starting small businesses? Dusting off the unfinished git repositories that never made it to prod? Share!