loveholidays has been adapting AI, moving to 50% AI commits in 6 months with a target of 60% by April 2026. The challenge for our engineers was to maintain and improve code health and DORA metrics while increasing productivity through AI-based innovation.

Six months in, we are forming learnings which support the importance of the agile values at the heart of the XP methodology. However, some of the practices, such as Pair Programming and TDD, are facing challenges. This session will explore our approach to measurement and the application of XP style guard rails, and the tension between by the book practice adherence and adaptations we see emerging.

We will share patterns that our teams are finding that work in the current agentic coding space and the non-negotiable measures which we use as guardrails in the agentic coding loop: emulating TDD and driving refactoring to high quality.

Learning outcomes: 1. Identify non-negotiable measures and guardrails that maintain code health 2 Measuring impact of AI adoption. 3. Understand emergent adaptations to TDD and Pair Programming that work effectively within an agentic coding loop.

Process/Mechanics: Style: lecture. Agenda: Case Study (5 mins): Review of the move from 0% to 50% AI commits and measurement approach. XP + agentic pattern adaption (10 mins): Explore friction points between traditional XP and agentic tools, focusing on pairing, TDD and refactoring. Guardrails (10 mins): Discussion of the approaches we take to inject Xp into the agentic loop and gitops & operational guardrails Q&A (5mins)

Room set-up: Standard theatre or classroom setup with a projector.

Target Audience: Software engineers, tech leads, and engineering managers focused on maintaining technical excellence while adopting AI.

Prerequisites: Basic knowledge of XP practices (TDD, Pairing, Refactoring) and familiarity with AI coding assistants.

www.linkedin.com/in/stuartcaborn

Stuart is a Distinguished Engineer at loveholidays, where he leads the Boost team helping engineering teams navigate complex technical and organisational decisions. He is fascinated by how company structures, ways of working, system designs, and AI shape each other. Recently he’s been collaborating with platform tech teams, exploring what makes AI work at the team level and how those patterns scale across technology.