For over five years, Centalion relied on Komgo’s platform to manage off-balance sheet instruments, yet teams remained tethered to manual data retrieval and status tracking across disparate systems. To address this, the company developed a cloud-based AI agent that connects to Komgo’s infrastructure via a Model Context Protocol (MCP) connector. This architecture enables the agent to execute structured workflows while keeping all sensitive logic and governance under Centalion’s direct control.

The impact on efficiency is significant. By automating the preparation of syndicated facility transactions, Centalion expects to recover over 60 working days of treasury capacity annually. According to Olivier Thevenon, Treasury Manager at Centalion, the goal was not to replace human decision-making but to remove operational friction. The agent interprets instructions and applies predefined business rules in seconds, leaving final approval firmly in the hands of the treasury team.

Baptiste Audren, Chief Revenue Officer at Komgo, noted that the deployment serves as a model for how corporations can build proprietary AI capabilities on top of established financial infrastructure. By avoiding point-to-point integration projects, the firms have created a scalable way to handle trade finance without sacrificing the accountability required in high-stakes commodity trading.