Source linked

ヴァージン・アトランティック、遺産コードを80%削減、30分でアプリをリリース

OpenAIのCodexを使用して、Virgin Atlanticは80%の古代コードを削減し、100%のユニットテストカバーを達成し、リスクの高いクリスマスに重要な欠陥を回避し、記録的な時間で新しいモバイルアプリを配信しました。

virgin atlanticcodexopenaineil lextfordrichard mastersrefactoring

Virgin Atlantic reduced a 100‑year‑old codebase by 80% in just 30 minutes, thanks to OpenAI’s Codex.

Speeding Up Legacy Refactoring

Engineering teams that had maintained legacy code for years now see a 78–80 % reduction in codebase size after a single Codex‑assisted session. Where refactoring once took two weeks, the airline reports a 30‑minute turnaround. In a recent sprint, a lead front‑end developer built a complete, working front‑end application from a Figma prototype in a week, with the backend stubbed out—so quickly that the Scrum master complained the backend tickets were still pending.

Zero‑Defect App Launch

The new mobile app, released in beta over Christmas and live weeks later, achieved near‑complete unit‑test coverage and zero P1 defects at launch. Codex helped the team hit the quality bar under deadline pressure, a feat rarely possible when teams must cut scope or testing to meet fixed dates. "Things don’t get delayed when we’re using Codex," VP of Digital Engineering Neil Letchford says.

Data‑Warehouse Prototyping

Codex also unblocks data‑engineering workflows. Analyst teams can prototype internal applications directly against the company’s data warehouse in a matter of hours or even a single workshop. "You can develop that data through to a prototype in a couple of hours," VP of Data and AI Richard Masters notes. The velocity gains extend beyond the app to migrations of databases onto the core warehouse.

Scaling the Momentum

With refactoring and prototyping now outpacing the rest of the delivery pipeline, Letchford’s next challenge is scaling Codex across the entire software development lifecycle. "How do we start scaling this up, not just in pockets, but across that whole software development lifecycle?" he asks. Masters adds that Codex is moving beyond pure engineers into a real tool for everyone.

The Virgin Atlantic case demonstrates that a generative‑coding agent can compress weeks of work into minutes, raise test coverage to 100 %, and eliminate critical defects—all while keeping the airline’s high‑stakes launch on schedule. As more enterprises adopt Codex, the boundary between human and machine‑assisted development will blur, reshaping how software is built and shipped.


Source: How Virgin Atlantic ships faster with Codex
Domain: openai.com

Read original source ->

External source stays available while the OJO article and comment thread stay local.

Comments load interactively on the live page.