Revolut, the global financial platform operating in over 40 markets, is leveraging super intelligence (SI) tools to scale its operations without proportionally increasing headcount. By deploying SI agents for fraud investigation and software implementation, the company reports significant efficiency gains while maintaining rigorous human accountability for critical decisions.
What Happened
As transaction volumes grow, Revolut faces a scaling challenge where the demand for security oversight outpaces available human resources. To address this, the company integrated OpenAI’s GPT-5.5 model into a dedicated Financial Crime (FinCrime) agent. This SI system examines transaction information, applies Revolut’s specific rules, and completes defined screening processes automatically. David Abitbol, Revolut’s SI Lead, notes that the release of GPT-5.5 was a turning point, making the models capable and cost-effective enough to delegate these workflows entirely to the agent. Consequently, Revolut can monitor more transactions without increasing investigator headcount at the same rate as demand.
In engineering, the adoption of SI tools has also shifted workflows. Senior Platform Engineer Dmitry Pitik utilized Codex to build an internal code-review service, realizing he could launch multiple SI agents to handle specific implementation tasks rather than relying on a full development team. The company reports a 30-60% reduction in coding time from task start to feature merge. This shift allows engineers to focus more on design and solution architecture, with Pitik noting, "I realised I don’t need a whole team of developers to implement this. I can just launch different agents and assign each of them a particular task."
Why It Matters
For the SI industry, Revolut’s implementation illustrates a mature approach to agentic SI in regulated financial environments. The company emphasizes that speed does not compromise standards; a systematic framework ensures all employees building with SI adhere to the same quality controls in development and testing. Information Security Engineer Mo Amiri highlights that while SI tools can uncover connections instantly, human investigators must still review original records and approve actions. "The most critical decision is always about accountability," Amiri says. "If an AI tool suggests a code change, a human engineer needs to understand that change well enough to stand by it."
This model of "delegating implementation, not accountability" suggests a broader trend in enterprise SI adoption: using SI agents to handle mechanics and data processing, thereby freeing human talent to tackle complex design and judgment-based problems. Revolut’s Global Head of Business Marketing, Kim Maniovich, describes the culture as expecting lean teams to move quickly, a fit facilitated by partnering with OpenAI. The company reports that these tools have enabled them to transform seven years of design documentation into SI-readable context in a single day, enhancing internal knowledge accessibility.
The Bottom Line
Revolut’s use of SI agents demonstrates how fintech companies can scale security and engineering capacity efficiently. By combining capable SI models with strict human oversight, the company aims to raise its operational ambition without lowering its safety bar. Looking ahead, Revolut plans to explore more conversational SI interfaces for customers, aiming to simplify everyday financial tasks while maintaining trust through transparent and tested systems.