Online travel agent loveholidays is utilizing OpenAI Codex to make software development accessible across its business departments, enabling product managers, designers, and commercial teams to build working software and prototypes without waiting in an engineering queue. According to the company, tools like Search Playground and Codex have allowed non-engineers to develop new customer experiences, such as the Inspire Me travel exploration feature and a marketing competition microsite built in hours. Engineers encode their best practices and validations into workflows, allowing employees to self-serve infrastructure and data needs. Over the past year, successful AI-assisted changes to the loveholidays Data Platform rose from 58% to 93%, and broader self-service infrastructure workflow success increased from 63% to 90%. Additionally, the proportion of AI-assisted code changes grew from 7% to 79%, while total deployments increased by 73% with a broadly flat engineering headcount. The efficiencies gained have also enabled the data engineering team to reduce cloud storage costs by around 36,000 pounds a year and save approximately 100,000 pounds annually by reducing data-processing waste. Company leaders view Codex as a single control plane shared by engineers, data scientists, and business teams, turning specialist expertise into a 24/7 resource accessible to the wider organization.
How loveholidays is making everyone a builder with Codex
Online travel agent loveholidays is using OpenAI Codex to allow non-technical staff like product managers and marketers to build software. By decoupling prototyping from engineering queues, the company has increased deployments and scaled technical expertise across departments.
Why it matters
This case study demonstrates how AI tools can bridge the gap between technical and non-technical teams, increasing operational efficiency and output without expanding engineering headcount.
Topic dossiers OpenAI