OpenAI: Revolut cuts coding time 30–60% with Codex, uses GPT-5.5 for financial crime checks
What happened
On November 16 OpenAI described how Revolut uses its models and Codex for software engineering and financial crime investigations. Coding time from task start to merged feature fell 30–60%, per the report. One engineer built an internal code review service using several Codex agents, and the team turned about seven years of design documents into AI-readable context in a single day. For financial crime, GPT-5.5 lets the FinCrime agent take on screening and investigation workflows. Revolut keeps one testing standard and requires people to understand code changes, check investigation evidence and approve actions.
Written by AI from the coverage · updated 3 days ago
Coverage
Follow the reports to see the story from different sides.
- OpenAI NewsHow Revolut raises ambition without lowering the bar
Revolut 使用 OpenAI 模型和 Codex 推进软件工程与金融犯罪调查,从任务开始到功能合并的编码时间减少 30–60%。一名工程师通过多个 Codex 智能体实现内部代码审查服务,团队还在一天内将约七年的设计文档转为 AI 可读上下文。GPT‑5.5 让 FinCrime 智能体能够承接筛查与调查流程,团队同时坚持统一测试标准,并要求人工理解代码变更、核查调查证据及批准行动。
Heat over time
Heat now 3·Comparable peak 5(Nov 17 15:00)·Comparable change over 24 hours -51%
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.