Growing Your Software Factory: Rules, Modules, and AI Leverage
Published on 01.04.2026
How to Grow Your Software Factory: Rules, Modules, and the Path to AI Leverage
TLDR: Luca Rossi of Refactoring argues that software engineering is shifting from a craft to a factory model, where the best teams systematically out-leverage everyone else through rules, modularity, and AI. The gap between elite and average teams is growing fast, and this piece is about how to close it, or widen it, depending on which side you're on.
Summary: Last month, Luca Rossi published "The Era of the Software Factory," co-authored with Rob Zuber, the CTO of CircleCI. That piece laid out a provocative thesis: we are no longer in the age of the lone genius programmer building cathedrals. Software engineering is becoming factory work. Not in a depressing, soul-crushing sense, but in the sense that systematic repeatability, modular thinking, and process discipline are becoming the dominant competitive advantages. This follow-up article, "How to Grow Your Software Factory," takes that argument and makes it operational.
The factory metaphor is a useful one, though it comes with baggage. When you say "factory," people picture assembly lines and interchangeable workers, which misses the point entirely. What Rossi means is closer to what the best manufacturing systems actually look like: lean, modular, with tight feedback loops and well-understood processes that allow skilled practitioners to focus on the work that matters rather than reinventing the wheel at every turn. That is a good thing. The question is how you build that kind of system for a software team.
The article zeroes in on two levers: rules and modules. Rules are the explicit, documented constraints that shape how code gets written, reviewed, and shipped. Not bureaucratic rules for their own sake, but the kind of lightweight guardrails that eliminate whole categories of decisions so your team can move faster. Modules are the modular building blocks, reusable components, shared libraries, internal platforms, that let a small team punch above its weight. Both of these become dramatically more powerful when you add AI into the mix, because AI thrives on well-structured, well-documented, rule-governed environments. Chaos is the enemy of AI leverage. Order is what unlocks it.
This connects to something I find genuinely interesting and somewhat underappreciated: the teams getting the most out of AI coding tools right now are not necessarily the ones with the most talented individuals. They are the ones with the cleanest codebases, the clearest conventions, and the most modular architectures. AI assistance scales with the quality of your engineering culture, which is a pretty interesting forcing function. It means all the boring stuff, consistent naming, good abstractions, documented decisions, now has a compounding return that it never quite had before.
The article is behind a paywall, so the full argument is not available here. But the preview makes it clear that the framing is squarely about the widening gap between elite and average teams. That gap is not primarily about individual talent. It is about systems. Teams that build the right factory infrastructure compound over time. Teams that do not fall further and further behind as AI multiplies every advantage that systematic teams already have.
Key takeaways:
- Software engineering is moving from craftsmanship to a factory model, where process discipline and modular thinking are the primary competitive advantages.
- The gap between elite and average teams is widening, driven by how well teams can systematize and leverage AI tools.
- Rules (explicit conventions and guardrails) and modules (reusable components and internal platforms) are the two core levers for building a high-leverage software factory.
- AI tools amplify the advantages of well-structured teams and well-organized codebases. Messy environments get less leverage, clean environments get dramatically more.
- This is a follow-up to "The Era of the Software Factory," co-written with Rob Zuber, CTO of CircleCI, and part of a longer arc of thinking about team productivity at scale.
Why do I care: From an architecture standpoint, this framing validates something many of us have felt for a while: the investment in internal tooling, shared component libraries, and explicit team conventions is not overhead. It is leverage. The "factory" critique can feel reductive when applied to creative engineering work, and I'd push back on any reading that treats software development as purely mechanical. But the core insight, that modular, rule-governed systems unlock AI leverage in ways that ad-hoc approaches cannot, is correct and important. Any senior architect thinking about how to set their team up for the next three to five years should be thinking hard about this. The teams building clean internal platforms today are the ones who will compound the fastest as AI capabilities continue to improve.