The Dutch bank reduced FX and credit trading system development from weeks to hours using AI, boosting large-ticket trade volumes by 50%.
ING has implemented AI-driven ‘vibe coding’ to develop electronic trading systems for its foreign exchange and credit businesses in hours instead of weeks. The bank confirmed the approach, using Anthropic’s Claude model, to create entire e-trading systems and real-time analytics dashboards with minimal developer input.
Previously, such projects required weeks of work by larger teams. ING’s electronic trading team, comprising around ten people, achieved a 50% increase in large-ticket trades using an in-house AI pricing model. The bank attributes its success to competing on innovation rather than headcount, as it lacks the resources of larger rivals.
The move highlights a broader industry shift toward AI-driven efficiency, with potential implications for productivity and risk management in banking.