loveholidays(在新窗口中打开)是一家领先的在线旅行社，业务覆盖八个欧洲市场，利用其技术每天处理60万亿种套餐组合，帮助数百万用户找到理想的假期。支撑这一规模的是公司多年来构建的技术平台，旨在让查找和预订假期变得更加快速和灵活。
如今，Codex正在改变谁能在这一平台上进行开发。
产品经理、设计师和商业团队正越来越多地直接参与loveholidays的代码库工作。团队可以制作客户体验原型、进行数据和基础设施变更，并将想法转化为可运行的软件，而无需每个请求都先进入工程队列。
“每个人都是构建者，”loveholidays工程主管Dmitri Lerko表示。“修改我们的应用程序、基础设施和部署代码不再是工程部门的专属活动。产品经理、设计师和商业利益相关者正在推动价值并执行部署。”
对于loveholidays的首席技术官Mike Jones来说，这是更大愿景的一部分：通过将技术与员工的专业知识相结合，构建公司所称的“旅行通用智能”——并通过AI让两者都更易于使用。
“在loveholidays，我们的平台愿景是构建旅行的通用智能。这需要将我们拥有的优秀技术与员工的专业知识结合起来——并通过AI和Codex将其民主化。”
——Mike Jones，loveholidays首席技术官
从想法到真实的客户体验
最明显的例子之一是Search Playground。
以前，公司其他部门的人如果有新客户体验的想法，需要说服工程团队优先制作原型。因此，每个实验都伴随着机会成本：用于测试一个想法的工程时间在其他地方就无法使用。
loveholidays希望打破这种依赖关系。
其工程师使用公司的设计系统、前端技术和Codex创建了Search Playground。它让公司各部门的人能够将想法转化为可运行的客户体验、收集反馈并测试其是否创造价值。
已有超过十种新的搜索体验通过Playground开发出来。其中大多数由非工程师构建，至少有三项已在loveholidays网站上运行。
另一个例子来自营销部门。在最近的《来自国外的薯片》活动中，团队希望建立一个互动微网站来收集竞赛参赛作品并分享度假灵感。以前，这需要依赖外部代理机构来设计和开发独立的数字体验，增加了成本和耗时。借助Codex和Search Playground，团队在数小时内自行构建了该体验，同时保持了loveholidays现有的设计系统。
“我们希望将尝试新想法的能力与实际工程时间解耦，”Lerko表示。结果不仅仅是更快的原型制作，而是更多想法有机会成为产品。
随时获取专业工程知识
同样的原则也适用于幕后工作。
loveholidays的数据平台和基础设施最初是为技术用户设计的。进行更改需要了解专业工具、代码仓库、版本控制和内部流程。当有人遇到困难时，需要专业工程师介入。
Codex为loveholidays提供了另一种扩展专业知识的方式。
工程团队将最佳实践、操作说明和验证规则编码到工作流程中，Codex可以引导其他用户完成这些流程。员工无需理解每个底层系统，只需专注于他们要完成的目标，而Codex则帮助提出变更建议、运行检查并引导其完成发布流程。
“我们将所有最佳实践编码化。我们将验证规则编码化，并根据现实变化持续改进，”Lerko表示。“我们的数据和基础设施工程师的专业知识通过Codex提供——因此任何自助使用基础设施或数据需求的人都能全天候获得这些专业知识。”
结果是可衡量的。
在过去一年中，loveholidays数据平台上AI辅助变更的成功率从58%上升到93%。与此同时，团队发现每项支持请求对应的数据平台变更数量增加了四倍。
在其更广泛的自助基础设施工作流程中，成功率从63%提升到90%。
这改变了两方面的工作。团队无需等待专家帮助即可推进，而工程师则花更少时间处理常规请求的故障排除，更多时间改进平台本身。
“你交给AI的工作越多，你的工作就越有提升。你的工作不再是拿到解决方案并实施它。你必须参与到业务问题中去。”
——Mike Jones，loveholidays首席技术官
更多软件，无需更多工程师
这一转变正在loveholidays的整个工程组织中显现。
一年前，其代码变更中约7%是AI辅助的。如今，这一数字为79%。
同期，部署量增加了73%，而工程人员数量基本持平。
对loveholidays而言，这种能力不是为了代码本身而生产更多代码。公司刻意以业务成果而非单纯的采用率来衡量AI。
“技术只是达到目的的手段，”Jones表示。“重点不在于技术本身，而在于它产生的影响。我们有意识地不仅让人们使用工具，还帮助他们解决业务问题——并衡量影响。”
部分影响已经体现在财务业绩上。
凭借更强的能力来处理那些此前机会成本过高的优化工作，loveholidays的数据工程团队已将云存储成本每年降低约36,000英镑，并通过减少数据处理浪费每年再节省约100,000英镑。
而更大的影响可能是团队现在认为什么是可能的。
“我们注意到，过去太难的事情现在变得平常了，”Lerko表示。“这意味着现在太难的事情将来也会变得平常。”
构建旅行通用智能
对loveholidays而言，Codex正日益成为员工与支撑公司的技术之间的通用界面。
“Codex正在成为一个单一控制平面——一个由工程师、数据科学家和业务部门共享的单一界面，”Lerko表示。“这蕴含着巨大的力量，因为你不再需要教每个人使用不同的工具。”
这指向了一种不同的软件构建模式：在这种模式中，专业知识可以在整个组织中共享，更多人可以将想法转化为可运行的产品，而工程师本身可以在问题解决链上更进一步。
对loveholidays而言，“旅行通用智能”最终意味着：将其构建的技术与员工的专长相结合，然后让这两者惠及更多业务环节。
而借助Codex，拥有想法的人与能够实现想法的人之间的界限正开始消失。
loveholidays(opens in a new window) is a leading online travel agent operating across eight European markets, using its technology to process 60 trillion package combinations every day to help millions of people find their perfect holiday. Behind that scale is a technology platform the company has spent years building to make finding and booking a holiday faster and more flexible.
Now, Codex is changing who gets to build on it.
Product managers, designers, and commercial teams are increasingly contributing directly to loveholidays’ codebases. Teams can prototype customer experiences, make data and infrastructure changes, and turn ideas into working software without every request first having to enter an engineering queue.
“Everybody is a builder,” says Dmitri Lerko, Head of Engineering at loveholidays. “Making changes to our applications, infrastructure, and deploying code is no longer an engineering-only activity. Product managers, designers, and commercial stakeholders are driving value and making deployments.”
For Mike Jones, CTO at loveholidays, that’s part of a bigger ambition: building what the company calls the general intelligence for travel by combining its technology with the expertise of its people—and making both more accessible through AI.
“At loveholidays, our platform vision is to build the general intelligence for travel. That’s bringing together the great technology we have with our people’s expertise—and democratising that with AI and Codex.”
—Mike Jones, CTO, loveholidays
From an idea to a live customer experience
One of the clearest examples is Search Playground.
Previously, someone elsewhere in the business with an idea for a new customer experience would need to persuade an engineering team to prioritise a prototype. Every experiment therefore came with an opportunity cost: engineering time spent testing one idea was engineering time unavailable elsewhere.
loveholidays wanted to break that dependency.
Its engineers created Search Playground using the company’s design system, frontend technologies, and Codex. It gives people across the business a way to turn an idea into a working customer experience, gather feedback, and test whether it delivers value.
More than ten new search experiences have already been developed through the Playground. Most were built by non-engineers, and at least three are now running on the loveholidays website.
Another came from marketing. For its recent Crisps from Abroad activation, the team wanted an interactive microsite to gather entries for a competition and share holiday inspiration. Previously, it would have relied on an external agency to design and develop a standalone digital experience, adding cost and time. Using Codex and Search Playground, the team built the experience itself in hours, while maintaining loveholidays’ existing design system.
“We wanted to decouple our ability to trial new ideas from actual engineering time,” says Lerko. The result isn’t simply faster prototyping. It means more ideas can earn the chance to become products.
Putting specialist engineering expertise on tap
The same principle applies behind the scenes.
loveholidays’ Data Platform and infrastructure were originally designed for technical users. Making changes required knowledge of specialist tools, repositories, source control, and internal processes. When someone got stuck, a specialist engineer had to step in.
Codex gives loveholidays another way to scale that expertise.
Engineering teams encode their best practices, instructions, and validations into workflows that Codex can guide other users through. Instead of needing to understand every underlying system, employees can focus on what they’re trying to accomplish while Codex helps propose a change, run checks, and guide it through the release process.
“We codify all the best practices. We codify validations, and continuously improve them as reality changes,” says Lerko. “The expertise of our data and infrastructure engineers is available through Codex—so anybody self-serving their infrastructure or data needs gets that expertise on tap, 24/7.”
The results are measurable.
Successful AI-assisted changes to loveholidays’ Data Platform have risen from 58% to 93% over the last year. At the same time, the team is seeing four times as many Data Platform changes for every support request.
Across its broader self-service infrastructure workflows, success has increased from 63% to 90%.
That changes the job on both sides. Teams can move without waiting for specialist help, while engineers spend less time troubleshooting routine requests and more time improving the platform itself.
“The more work you can hand off to AI, the more your job elevates. Your job isn’t to be handed a solution and implement it anymore. You have to get involved in the business problem.”
—Mike Jones, CTO, loveholidays
More software, without more engineers
That shift is showing up across loveholidays’ engineering organisation.
A year ago, around 7% of its code changes were AI-assisted. Today, that figure is 79%.
Over the same period, deployments have increased 73%, while engineering headcount has remained broadly flat.
For loveholidays, that capacity isn’t about producing more code for its own sake. The company deliberately measures AI against business outcomes rather than adoption alone.
“Technology is just a means to an end,” says Jones. “It’s not about the technology itself; it’s about the impact it has. We’re intentional about not just giving people access to tools, but helping them solve business problems—and measuring the impact.”
Some of that impact is already visible in the bottom line.
With more capacity to tackle optimisation work that previously carried too high an opportunity cost, loveholidays’ Data Engineering team has reduced cloud storage costs by around £36,000 a year and is saving approximately another £100,000 annually by reducing data-processing waste.
And the bigger effect may be what teams now consider possible.
“We’re noticing that what used to be too hard is now ordinary,” says Lerko. “The implication is that what’s too hard now will become more ordinary.”
Building the general intelligence for travel
For loveholidays, Codex is increasingly becoming a common interface between its people and the technology underpinning the company.
“Codex is becoming a single control plane—a single interface shared by engineers, data scientists and the business,” says Lerko. “There’s a lot of power in that, because you no longer need to teach everyone a different tool.”
That points toward a different model for building software: one in which specialist expertise can be made available across an organisation, more people can turn ideas into working products, and engineers themselves can move further up the problem-solving stack.
For loveholidays, that’s ultimately what “general intelligence for travel” means: combining the technology it has built with the expertise of its people, then making both available to more of the business.
And with Codex, the distinction between the people who have ideas and the people who can build them is beginning to disappear.
本文内容采集自官方网站，排版和翻译可能与原页面存在差异。
阅读官方全文