Loveholidays is quietly pulling off something most companies only talk about: turning non-engineers into actual contributors to production codebases. The online travel agent, which processes 60 trillion package combinations daily across eight European markets, has seen AI-assisted code changes jump from 7% to 79% in a single year. Deployment frequency rose 73% without adding a single engineer.
The mechanism is Codex, OpenAI’s coding model, but the insight is organizational. Loveholidays built a tool called Search Playground that wraps the company’s design system and frontend stack around Codex, letting product managers, designers, and marketing teams spin up working prototypes without queuing behind engineering priorities. More than ten new search experiences have come out of it, most built by non-engineers, and at least three are live on the production site. One, called Inspire Me, helps travelers browse trip types from beach breaks to foodie escapes. Another, a competition microsite for the Crisps from Abroad campaign, went from concept to working experience in hours—work that previously would have been outsourced to an agency.
“Everybody is a builder,” says Dmitri Lerko, Head of Engineering. That’s not marketing fluff. The company’s Data Platform tells the same story: successful AI-assisted changes climbed from 58% to 93% over the last year, while the platform now handles four times as many changes per support request. Self-service infrastructure workflows improved from 63% to 90% success.
The shift isn’t just about speed. It’s about who gets to propose ideas. Previously, a commercial team with a customer experience concept had to persuade an engineering manager to prioritize it. Every prototype carried an opportunity cost. Now that bottleneck is dissolving—which forces a harder question: when everyone can build, what do engineers actually do? At loveholidays, the answer so far is encoding best practices into Codex-guided workflows and tackling the genuinely difficult problems that still require deep technical judgment.
