
What happened
Revolut's AI Lead David Abitbol says OpenAI's Codex delivered a 30-60% reduction in coding time, and a GPT‑5.5-powered FinCrime agent now runs screening and investigation workflows automatically.
Why it matters
That lets Revolut screen more transactions without scaling investigators one-for-one, and it frees engineers to spend more time on design and review rather than implementation.
WHO IT HITSCompliance and financial-crime investigators at banks and fintechs stand to see their manual screening work partly automated, while engineering teams may shift time from implementation toward design and review.
Summaries like this, every morning.
Revolut's account, published by OpenAI, positions the fintech's AI use around two pressure points: transaction volume that outpaces human review capacity and engineering time spent on implementation rather than design. The reported 30–60% reduction in coding time is framed less as a cost cut than as a shift in where engineering expertise is applied—toward system design, planning, and review. The FinCrime agent, powered by OpenAI models, is described as examining information and applying Revolut's rules to complete screening and investigation steps automatically, which the company says lets it monitor more transactions without increasing investigator headcount at the same pace as demand.
A consistent theme is that delegation stops short of accountability. Information Security Engineer Mo Amiri describes tracing what information a system sees, where it goes, what tools it can reach, and what happens when something fails, while engineers remain responsible for code they produce. Internal examples—an engineer delegating implementation to multiple Codex agents, and seven years of design documentation turned into AI-readable context in a single day—illustrate the kind of leverage Revolut says it is getting from the models.
Looking ahead, Revolut points to a more conversational customer interface for tasks like blocking a card or managing a subscription. The company's framing is that as building gets easier, the differentiator becomes knowing what to build and keeping the judgment and standards to do it well.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Innovation BASE Hokkaido updated AI:Promo's image editing with OpenAI's GPT-Image-2.6 Sunburst, announced Nove…

Softcreate's Yamaguchi published a four-phase guide for Microsoft 365 Copilot: prepare a daily-report prompt…

SUGENA will hold a free Zoom webinar on December 17, 2026, capped at 100 pre-registered attendees, covering 20…

Anfini will hold a seminar titled “Human Capital AI Management” at ResorTech EXPO 2026 in Okinawa on November…

ASUS said it is expanding production of NVIDIA Vera Rubin NVL72 systems in Q4 to meet rising customer demand…

At Semafor’s Silicon Valley & The World, Mubadala CEO Khaldoon Al Mubarak said a "winter" correction is coming…
