AI and Software Craftsmanship: Why Quality Practices Matter More Than Ever

Published on February 5, 2025

4 min read

At a recent lightning talk at Flutter Belgium, we shared a crucial perspective on AI in software development: AI isn’t replacing software craftsmanship - it’s amplifying it.

AI as a Skill Magnifier

The core insight is simple but profound: AI acts as a mirror and magnifier of your existing development practices. For junior developers without established methodologies, AI can accelerate poor practices. For experienced teams with strong fundamentals, AI becomes a powerful accelerator.

This creates an interesting challenge for the industry. While some executives see AI as a way to bypass the need for skilled developers, and some developers proudly resist using it altogether, high-performing teams are quietly integrating AI with proven practices.

The Three-Step Approach

Rather than jumping straight to AI code generation, we advocate for a more strategic approach:

Key Principles

Step 1

Clarify the Work

Clarify the Work

Start with requirements, user stories, and acceptance criteria. AI excels at spotting gaps and ambiguity in specifications. Use it to strengthen your understanding before writing any code. This foundational step ensures that both human developers and AI tools have a clear target to work toward.

Step 2

Build Guardrails

Build Guardrails

Use AI to make mistakes harder by establishing: • Idiomatic code standards • Clear naming conventions • Consistent structure • Policy and compliance checks The question shifts from "does it work?" to "is it fit to live here?"

Step 3

Then Accelerate

Then Accelerate

Only after establishing quality foundations should you use AI to accelerate delivery. Speed becomes the outcome, not the goal. With clear requirements and solid guardrails in place, AI can safely accelerate your development process without compromising quality.

Why Software Craftsmanship Still Matters

The practices that define high-performing teams (identified by DORA metrics) remain essential in an AI-enhanced world:

-Test-Driven Development (TDD): Provides the feedback loops that guide AI-generated code
-Trunk-Based Development: Enables continuous integration when AI accelerates output
-Continuous Delivery: Ensures quality gates catch issues before production

These practices don’t just apply to human developers. Agentic frameworks and AI coding assistants benefit from the same disciplines that make human teams successful.

The Real Challenge for Junior Developers

The challenge for newcomers to development has shifted dramatically. They now need to understand architectural principles, design patterns, and quality practices to effectively direct and validate AI output - skills that previously came from years of hands-on coding experience.

This doesn’t mean we don’t need junior developers. It means their learning path must evolve to emphasize these foundational concepts earlier.

Looking Forward

As more people gain access to AI coding tools, software quality becomes not just a technical concern but a competitive differentiator. Teams that establish strong practices before accelerating with AI will create better systems, faster.

The question isn’t whether to use AI, but how to use it responsibly within a framework of proven software craftsmanship.

“AI is really a magnifier of skills of an individual or of a team. Software crafting skills have been proven as underlying principles under high performance teams.”

This talk was inspired by posts from Andrea Laforgia and Bryan Finster on software craftsmanship and AI.

Ready to leverage AI effectively in your development process?