Holiday Extras Codex adoption has reshaped how the travel-services firm builds apps and runs product development, enabling staff across the company to turn customer and operational knowledge into working software. By combining OpenAI’s Codex with organisation-wide ChatGPT Enterprise access, Holiday Extras shortened development timelines, made releases more frequent, increased booking conversion and reduced app errors.
The company had historically maintained mobile applications on an annual release cadence. After rolling out ChatGPT Enterprise broadly, Holiday Extras introduced Codex to assist in rebuilding both customer-facing applications and internal systems. The development model shifted from a mostly engineer-driven process to a collaborative workflow in which designers, product teams and operational experts used Codex to translate user insights into code and to create small, documented packages that served as reusable building blocks.
Holiday Extras reports a striking acceleration in a core rebuild: a project that would previously have taken more than six months produced a fully working version in about two weeks. Engineers retained responsibility for architecture, APIs, security and development workflows, while broader teams engaged more directly on feature design and implementation. The result was not only speed but also a modular approach that allowed teams to iterate on discrete components without rebuilding entire systems.
The rebuilt application delivered measurable improvements in user outcomes and engineering cadence. Open-to-book conversion on the main app increased by 32.5% after the rebuild, and the app’s error rate fell from 19% to 9%. Release frequency moved from a single major annual update to weekly releases in the first month following the rebuild, with each release containing multiple new features. Smaller product lines saw similar benefits: an internal product called Countdown went from adding a feature roughly every six months to multiple feature additions per week.
Faster iteration enabled Holiday Extras to respond to customer feedback with unprecedented speed. The company cites a case where a customer’s one-star review was transformed into a five-star rating after a prompt update addressed a missing list view. Rapid fixes and iterative improvements made it possible for product teams to treat customer feedback as immediate input to the development cycle rather than items queued for the next annual release.
The programme extended beyond product development into operations. Holiday Extras implemented an automated system that converts airport-transfer customer chats into actionable tickets, surfacing issues across its supplier network. Leaders also point to internal examples of non-engineering staff reshaping systems: a long-serving accountant rebuilt finance tools based on operational insight, showing how domain experts can directly influence tooling when given access and support.
Leaders at Holiday Extras emphasise that the initiative was designed to empower employees who know customers and operations to act on that expertise. Staff were encouraged to experiment with AI tools while engineers established guardrails, packaged components and security controls to keep experimentation practical and safe. That blend of empowerment and governance allowed the organisation to scale AI-assisted work without compromising core engineering responsibilities.
Holiday Extras describes OpenAI’s tools as capable of exploring ideas, reasoning about complex problems, proposing plans and executing work with minimal supervision. ChatGPT Enterprise provided organisation-wide support, while OpenAI models also power the app’s personalised customer experience. The company also used a coordination tool called dots to provide proactive performance reporting after launches. Holiday Extras leadership values both the model quality and the working relationship with the OpenAI team, describing the partnership as responsive to feedback during integration.
With accelerated development cycles and broader internal use of AI, Holiday Extras is reconsidering the scope of opportunities it can pursue. The app rebuild has become an example of how enabling more employees to act on their expertise can change product roadmaps and operational strategy, allowing the business to prioritise initiatives that might previously have been out of scope. The company intends to continue expanding systems administered through AI, with employees directing outcomes and leadership supporting ongoing experimentation and adoption.
Holiday Extras’ experience illustrates how pairing coding models like Codex with enterprise-wide conversational AI can shift the balance of who builds software and how quickly organisations can iterate. The results at Holiday Extras—faster releases, higher conversion and fewer errors—offer a concrete case study for other companies considering how to integrate AI into product and operational workflows.
Source: Read the original source
Leave a Reply