Quick Overview
This case study interview produced by OpenAI features loveholidays CTO Mike Jones and Head of Engineering Dmitri Lerko. The executives discuss how the European online travel agency integrated OpenAI Codex across engineering workflows and non-technical departments to accelerate development cycles.
Key Points
- 1.loveholidays has adopted AI-assisted coding across roughly 80 percent of its codebase within one year.
- 2.Engineers built an internal portal called Search Playground that links front-end design systems to Codex, enabling non-engineers to create novel search experiences.
- 3.Deployment frequency increased by approximately 73 percent while maintaining broadly flat headcount across the engineering organisation.
- 4.Data changes doubled while inbound internal support requests dropped by half due to codified engineering best practices and validations.
- 5.Codifying infrastructure rules into Codex allows product managers, designers, and commercial stakeholders to self-serve and deploy changes autonomously.
Summary
CTO Mike Jones introduces the operational scope of loveholidays, an online travel agency operating in eight European countries and evaluating up to 60 trillion package combinations daily. Jones argues that delegating standard coding and implementation tasks to AI elevates an engineer's role, shifting their focus toward understanding and solving fundamental business problems. The company's platform objective centers on combining travel expertise with automated intelligence, using Codex as the unifying mechanism to democratise technical execution across departments.
Head of Engineering Dmitri Lerko outlines how Codex operates as a shared control plane across software engineers, data scientists, and commercial business units. By codifying architecture standards, best practices, and data validation rules, loveholidays embeds the domain knowledge of senior infrastructure and data engineers directly into tooling. This codified knowledge base is continually updated when assumptions fail or system requirements evolve, providing every team with on-demand guidance and self-service capabilities around the clock.
To decouple experimentation from dedicated engineering resources, loveholidays built an internal development environment called Search Playground. The portal integrates the company's existing front-end technologies and design system with an interface powered by Codex. Non-technical employees have used this portal to construct more than ten novel search experiences. These functional web prototypes are shared directly with external customers for feedback and deployed into live experimentation on the main website.
Over the span of a single year, AI-assisted coding increased from near zero to roughly 80 percent of all code created at loveholidays. Jones reports that software deployment frequency rose by approximately 73 percent over this timeframe without requiring an increase in engineering headcount. Simultaneously, the volume of data modifications doubled across the organization while incoming support tickets for data teams fell by 50 percent, allowing data engineers to focus on platform architecture rather than routine operational requests.
Jones and Lerko conclude that technology serves strictly as a tool for measurable business outcomes rather than an end in itself. By lowering technical barriers, the organisation has shifted code deployment and application changes from an engineering-only function to a shared practice that includes product managers, designers, and commercial business stakeholders.
Vision and Role of Codex
CTO Mike Jones describes the company's objective to build general intelligence for travel across eight European markets. Codex acts as a central mechanism to democratise technical work, freeing staff from pure implementation tasks so they can focus on solving underlying business problems.
Codifying Infrastructure and Knowledge
Head of Engineering Dmitri Lerko explains that Codex functions as a single shared control plane across engineers, data scientists, and business teams. Best practices and validation rules from infrastructure and data specialists are codified into the system, offering continuous 24/7 self-service guidance across the company.
The Search Playground Portal
loveholidays created an internal tool called Search Playground that pairs company design systems and front-end code with Codex. This platform has allowed non-engineers to build over ten new search prototypes, run real user experiments, and collect customer feedback without consuming core engineering time.
Productivity Metrics and Organisational Impact
AI assistance in code generation rose from zero to roughly 80 percent in one year, driving a 73 percent increase in deploys with static headcount. Autonomous data workflows doubled data modifications while cutting support tickets in half, expanding code deployment responsibilities to product managers and commercial teams.
The Bottom Line
The video demonstrates how loveholidays integrated OpenAI Codex into internal developer tooling to increase deployment frequency and expand product prototyping beyond the core engineering team. It establishes concrete organizational metrics, including an 80 percent coding adoption rate, a 73 percent increase in deployments, and a halving of support tickets. The discussion leaves open how loveholidays manages code quality governance and long-term maintenance as non-technical stakeholders continue shipping production changes.
FAQ
What is OpenAI Codex and how is it used across loveholidays?
OpenAI Codex is an AI model that generates code and powers developer tooling, serving as a unified control plane at loveholidays to help engineers and non-technical staff build, validate, and deploy applications.
How does the Search Playground portal enable non-engineers at loveholidays to build search experiences?
Search Playground connects company front-end design systems with Codex, allowing non-engineers to prototype functional websites, gather customer feedback, and run live search experiments without taking up core engineering hours.
How much has AI-assisted coding adoption grown at loveholidays over the past year?
AI-assisted coding expanded from essentially zero to approximately 80 percent of all code written across the company in a single year.
What measurable improvements in deployments and support requests did loveholidays achieve using Codex?
Deployments increased by about 73 percent with a stable headcount, while data changes doubled and internal support requests dropped by half.
How does codifying infrastructure and data engineering expertise benefit employees across loveholidays?
Codifying rules and validations ensures that best practices from senior data and infrastructure engineers are available 24/7, allowing other employees to self-serve their data needs autonomously.
Worth watching for
Engineering leaders, software developers, and product managers interested in enterprise adoption of AI coding assistants and internal developer platform tooling.
- openai
- codex
- loveholidays
- software-engineering
- developer-productivity
- ai-coding