It Doesn’t Help To Push AI Into A Crappy Process

2026-04-29 · Source: Modern Software Engineering · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, long

Summary

Emily Batche, creator of Salmon Coaching, highlights that integrating AI tools into software development processes yields vastly different results depending on the existing methodology. For teams with a "coding-driven development" (CDD) approach, AI often accelerates existing inefficiencies, leading to a reported 30-50% speed-up in coding tasks but without improving code quality, design, or testing practices. This "code generation fire hose" can exacerbate downstream bottlenecks. Conversely, teams employing a "test-driven development" (TDD) process leverage AI more effectively. AI assists in understanding problems, sketching development plans, building small, testable components, and safely refactoring code while maintaining test integrity. This TDD-AI synergy results in higher quality code, more frequent commits (cycle times under 10 minutes), and improved design, demonstrating that a robust underlying process is crucial for realizing AI's full potential.

Key takeaway

For Software Engineers or ML Engineers aiming to integrate AI into their workflow, prioritize refining your core development processes first. If your team uses coding-driven development, adding AI will likely only accelerate existing issues like poor design and insufficient testing. Instead, adopt a Test-Driven Development (TDD) mindset; AI tools can then significantly enhance problem understanding, test list generation, and safe refactoring, leading to higher quality code and faster, more reliable delivery cycles.

Key insights

AI amplifies existing software development process quality; good processes benefit, poor ones worsen.

Principles

• A solid engineering process is foundational for AI tool effectiveness.
• AI integration with TDD improves code quality and development cycle time.
• Speeding up a flawed process with AI exacerbates existing problems.

Method

Integrate AI into TDD by using it to understand problems, sketch test plans, build small components, and refactor code, ensuring frequent, high-quality commits.

In practice

• Use AI to generate examples and validate problem understanding.
• Employ AI for slicing problems and creating test lists in TDD.
• Prompt AI to refactor code while maintaining test integrity.

Topics

• AI in Software Development
• Test-Driven Development
• Coding-Driven Development
• Software Process Improvement
• Code Refactoring
• Development Cycle Time

Best for: Software Engineer, Machine Learning Engineer, Director of AI/ML

Related on AIssential

• The hidden pearls of TDD — TDD is a foundational discipline that enhances software quality, design, and AI agent effectiveness through structured feedback and complexity prevention.
• Beyond Code Review — Specification-Driven Development shifts focus from code review to continuous verification against evolving specifications.
• My AI Coding Workflow — The author's AI coding workflow leverages a "Guides, Gates, and Guards" model, shifting to Codex for efficiency and cost reduction.
• Fragments: April 14 — AI's transformative power necessitates adapting established software engineering principles like TDD and abstraction, while designing for AI restraint.
• Cut Your Debugging Time By 80% with AI Skills And This Workflow — Combining TDD, vertical slicing, and AI tools significantly boosts code quality and reduces debugging.
• Cycles of disruption in the tech industry: with software pioneers Kent Beck & Martin Fowler — AI's rapid adoption and transformative potential echo past tech shifts, demanding a balance of skepticism and curiosity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Modern Software Engineering.