Revolut uses OpenAI models and Codex for engineering and fraud
Original title: How Revolut raises ambition without lowering the bar
Revolut says GPT-5.5 made a FinCrime agent cheap enough to run screening without matching investigator growth.
OpenAI published a customer story on Revolut’s use of ChatGPT, the API and Codex across engineering and financial crime. Revolut says Codex cut coding time from task start to feature merge by 30–60% and saved one to two days on security code review. AI lead David Abitbol says GPT-5.5 was the point at which models could take screening and investigation workflows. Engineers still own the code and Revolut says it does not blindly trust AI output.
Unconfirmed: Time savings and workflow claims come from Revolut staff in OpenAI’s customer story.
- Revolut cites more than 40 markets and 800,000 businesses.
- One cited report says over half of adults worldwide reported a scam encounter in 2025.
- Staff turned seven years of design documentation into AI-readable context in one day.
- A platform engineer used separate Codex agents as an implementation team for an internal review service.
- Revolut applies shared standards for AI development, A/B testing and business-metric prediction.
From the original article
9 min read at openai.comDiscover how Revolut uses OpenAI models and Codex to cut coding time by 30–60%, scale financial crime investigations, and innovate without compromising security.
openai.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.