Two People, 560K Lines, 5 Months
What AI-augmented enterprise software development actually looks like - not a theoretical answer, but a practical one with real numbers, real mistakes, and an actual product running in production.
Introduction
There is a conversation happening in every software company right now, whether they admit it or not. Can a small team, armed with the right tools and the right approach, deliver what used to take a department (or several)? And if so, what does that mean for how we build software, how we staff teams, and how we think about productivity in general?
In many ways it's a scary topic and some keep their fingers crossed that the real answer is "No". And we get that, we really do. On one side it's amazing what "it" gives you a chance to do, on the other one it's concerning how it will affect your business and your livelihood. But we will concentrate on going forward, learning this new skillset and challenging ourselves to be the builder and thinker, because this is the tool that gives you the chance to become "your own".
So in this article we will address those questions and show how it really is, on one of our flagship systems. Not a theoretical answer - a practical one, with real numbers, real mistakes, and a real product that's running in production as you read this.
Mind you, we come from many years in IT, so that should be taken into consideration when reading through our experience and conclusions.
What is InTakt?
InTakt is an operations platform for small and mid-sized manufacturers, installers, and field service companies. The kind of companies that outgrew Excel two years ago but can't justify a six-figure SAP implementation and a year-long onboarding process. We built InTakt to sit in that gap - real warehouse management, work orders, procurement, sales, and a mobile app that streamlines day-to-day warehouse operations.
Now here are the numbers, and this is where it gets interesting.
The system has 145 database entities with full relational integrity, not a flat table dump with creative column naming. It runs on ASP.NET Core 8 with PostgreSQL, has a Flutter mobile app for Android and iOS, a web admin portal, Docker-based deployment with CI/CD, and centralized logging and monitoring/metrics. The total codebase sits at around 560,000 lines of code.
Two people built this. In five months.
I know what you're thinking - and no, it's not a prototype, it's not a demo, and it's not held together by duct tape and optimism. It's a production system with real tenants, real data, and a deployment pipeline that ships daily. The natural question is "how?" and that is exactly what the rest of this article is about. But before we get into the methodology, the tooling, and the honest parts where things went wrong, we want to be clear about one thing. This is not a story about AI replacing developers. This is a story about two experienced professionals who found a way to multiply their output by an order of magnitude, without compromising on the things that actually matter in production software.
If you're here for a prompt engineering tutorial, this blog is not it. If you're here to understand how a small team with the right approach can build systems that would normally require five to seven people - keep reading.
Ready to build something meaningful?
Let's talk about your project. Tell us what you need, we'll respond within 24 hours.