Tech • AI • Robotics
What Codex Unlocks for loveholidays
5/10TL;DR
loveholidays says AI-assisted coding now produces about 80% of its code, helping the online travel company speed up releases, widen software-building beyond engineers and tie technical work more directly to business problems.
KEY POINTS
AI moves software work up the value chain
At loveholidays, the role of technical staff is shifting from simply implementing predefined solutions to working more closely on business problems. The company’s view is that the more repetitive execution can be delegated to AI, the more employees can focus on deciding what should be built and why.
A major European travel platform
loveholidays operates in eight European countries as a leading online travel agent. Its technology processes roughly 60 trillion package combinations each day to help customers find travel options, giving the company a large-scale search and data environment in which AI tools can be applied.
A push toward “general intelligence for travel”
The company’s platform strategy is to build what it describes as general intelligence for travel, combining internal technology with human expertise. In that model, Codex is being used as a way to spread technical capability across the organization rather than confining it to specialist teams.
Codex as a shared control plane
Codex is being positioned as a single interface shared by engineers, data scientists and business teams. Instead of training each group on separate tools, the company is embedding best practices, validations and updated assumptions into one system, creating a common layer for technical and operational work.
Specialist knowledge is being codified
Expertise from data engineers and cloud infrastructure engineers is being turned into reusable rules and workflows inside the platform. That means employees serving their own infrastructure or data needs can access specialist guidance continuously, while expert teams spend less time handling routine requests.
Non-engineers are building search products
Engineers created a portal called Search Playground, built on the company’s design system and front-end technologies and connected to Codex. Through it, staff can turn ideas into working websites, share them with external customers, collect feedback and launch live experiments.
More than 10 new search experiences were created
The company says it has developed more than 10 novel search experiences through this approach, with the majority created by non-engineers. That is meant to reduce dependence on scarce engineering time and make it easier to test new commercial or product concepts quickly.
Coding output and deployment speed increased sharply
Over the past year, AI-assisted coding rose from almost zero to about 80% of all code produced. During the same period, deployments increased by about 73% while headcount stayed broadly flat, suggesting the company is using AI to raise output without a comparable increase in staffing.
Data work expanded while support demand fell
loveholidays says the amount of changes made to its data roughly doubled, yet support requests dropped by about half. The result is greater autonomy for data-platform users and more time for infrastructure engineers to improve the platform and tackle new technical problems.
Software delivery is no longer engineering-only
The company says changes to applications, infrastructure and deployed code are no longer limited to engineering teams. Product managers, designers and commercial stakeholders are increasingly able to make changes directly, reflecting a broader shift toward treating software creation as a cross-functional business capability.
CONCLUSION
For loveholidays, the central promise of AI is not only faster coding but a wider redistribution of technical power across the company. The key test will be whether that broader access continues to improve business outcomes while preserving quality, controls and reliability.
Explain this