Continuous Delivery in a World of Constant Change • Abby Bangser & Dave Farley • GOTO 2025

2026-05-25 · Source: GOTO Conferences · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, extended

Summary

Continuous Delivery (CD) is defined as ensuring software is always in a releasable state, achieved through small, verifiable steps and constant feedback. This approach, fundamental to scientific and engineering problem-solving, guarantees better outcomes in software development. In the context of AI and Large Language Models (LLMs), these established XP-style practices, such as small, testable changes and the ability to revert, become even more critical. While AI offers potential for abstraction, it often struggles with incrementalism and precision, leading to "messy" code. The speakers emphasize that AI programming must facilitate problem decomposition, human communication, and incremental progress. Behavior-Driven Development (BDD) and Test-Driven Development (TDD) are proposed as essential methods to precisely specify desired outcomes, allowing for verification of AI-generated code. The discussion also highlights the ongoing challenge of human verification for AI output and the organizational "people problems" that AI cannot solve, such as coupling between teams. A recent Dora report indicates experienced programmers are better at steering AI, raising concerns about the development of junior talent.

Key takeaway

For development teams integrating AI into their workflows, prioritize established Continuous Delivery principles like small, verifiable steps and testable changes. AI's propensity for non-deterministic output and large code rewrites makes precise specification via Behavior-Driven Development (BDD) or Test-Driven Development (TDD) crucial. You must actively "trust but verify" AI-generated code, ensuring human oversight and robust testing to maintain quality, control incremental progress, and avoid the pitfalls of unconstrained automation.

Key insights

Continuous Delivery's incremental, verifiable steps are foundational for robust software development, especially with AI.

Principles

• Software must always be in a releasable state.
• Solve complex problems through small, verifiable steps.
• Programming aids problem decomposition, communication, and incremental progress.

Method

Employ Behavior-Driven Development (BDD) to precisely specify desired outcomes, then execute these specifications as tests against AI-generated code to ensure accurate fulfillment.

In practice

• Adopt XP-style practices for AI-assisted development.
• Use BDD specifications to verify AI-generated code.
• Cultivate human oversight for AI code quality.

Topics

• Continuous Delivery
• AI Development
• Behavior-Driven Development
• Test-Driven Development
• eXtreme Programming
• Software Quality Assurance

Best for: Software Engineer, DevOps Engineer, AI Engineer

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.
• Spec Driven Development isn't Waterfall — Specification Driven Development iterates on living specifications, enabling autonomous AI agents to accelerate software delivery.
• CI/CD Is Dead, Agents Need Continuous Compute and Computers — Hugo Santos and Madison Faulkner — Agentic software's high-volume, continuous code generation necessitates a shift from traditional CI/CD to continuous compute.
• Beyond Code Review — Specification-Driven Development shifts focus from code review to continuous verification against evolving specifications.
• Fragments: April 14 — AI's transformative power necessitates adapting established software engineering principles like TDD and abstraction, while designing for AI restraint.
• It Doesn’t Help To Push AI Into A Crappy Process — AI amplifies existing software development process quality; good processes benefit, poor ones worsen.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by GOTO Conferences.