A New Era for Trade Finance Automation

In a significant advancement for corporate treasury operations, Centalion Group and Komgo have announced the live deployment of an AI Treasury Agent. This initiative represents a practical application of agentic artificial intelligence within the complex domain of commodity trading and trade finance. Announced on Oct. 14, 2026, in Geneva, the deployment allows treasury professionals to execute facility and trade finance instrument workflows using natural language, while ensuring that all existing control and approval steps remain fully intact.

Centalion Group, recognized as one of the world's largest independent commodity trading companies, has partnered with Komgo, a leading digital trade finance platform provider, to bridge the gap between advanced language models and structured financial infrastructure. The deployment marks a shift from passive automation to active, agentic workflows that operate securely within a corporate's governed cloud environment.

Overcoming the Friction of Multi-System Workflows

Prior to this deployment, Centalion had established a long-term operational foundation with Komgo, relying on the Konsole platform for over five years. Konsole served as the central platform for managing Centalion's off-balance sheet instruments across its various banking counterparties.

However, even with robust digital platforms in place, treasury teams frequently encountered operational friction. Professionals were required to manually navigate between disparate systems to perform routine tasks, such as:

  • Retrieving critical transaction data across platforms
  • Validating information against internal business rules
  • Chasing status updates from banking counterparties

This manual coordination created administrative bottlenecks, consuming valuable time that could otherwise be allocated to strategic financial management. To address this, Centalion sought a solution that could automate these repetitive tasks without compromising security or compliance.

Technical Architecture: The Model Context Protocol Connector

Rather than adopting a generic chatbot or a rigid third-party software solution, Centalion developed and governed the AI Treasury Agent within its own secure cloud environment. This architecture combines a large language model with structured treasury workflows and facility-specific business rules.

The integration between Centalion's proprietary AI environment and Komgo's trade finance infrastructure is powered by Komgo's Model Context Protocol (MCP) connector. This technical approach offers several distinct advantages:

  • Control and Orchestration: Centalion retains complete ownership and control of its AI models and orchestration logic.
  • Secure Interaction: The agent securely interacts with Komgo's trade finance infrastructure without exposing sensitive internal rules.
  • No Point-to-Point Integration: The MCP connector eliminates the need for complex, costly, and time-consuming point-to-point integration projects.

By utilizing the MCP connector, the companies have demonstrated a blueprint for how corporate treasuries can leverage external financial networks while maintaining strict internal data sovereignty.

Quantifiable Operational Impact

The first deployment of the AI Treasury Agent focuses on Centalion's off-balance sheet instrument syndicated facility. The operational results collected at the transaction level demonstrate immediate efficiency gains:

Issuance Processing
55

The time required to process issuances has been reduced by around 55%.

Amendment Processing
80

The time required to process amendments has been cut by around 80%.

When these capabilities are extended across Centalion's entire trade finance activity, the company projects significant long-term capacity savings. The automation is expected to free more than 60 working days of treasury capacity each year. This newly reclaimed time will be redirected toward high-value activities, including:

  • Risk oversight and mitigation
  • Facility optimization
  • Strategic financing decisions

Maintaining Governance and Human Accountability

A core philosophy of the deployment is the preservation of human oversight. The AI Treasury Agent is designed to assist, not replace, human decision-makers. It interprets natural language instructions, applies pre-defined business rules, and prepares transactions in seconds, but the final approval and execution authority remains strictly with the treasury team.

Olivier Thevenon, Treasury Manager at Centalion, emphasized this balance, stating that the goal was not to build another chatbot, but to fundamentally change how the treasury team interacts with technical infrastructure. Thevenon noted that the agent can understand an instruction, apply rules, and prepare a transaction in seconds while the decision remains with the people in the treasury team. For Centalion, this deployment demonstrates the real potential of agentic AI in a commodity trader's treasury department: removing friction without removing accountability.

Baptiste Audren, Chief Revenue Officer at Komgo, highlighted the broader industry implications of this approach. Audren pointed out that the collaboration shows how corporate treasury teams can build and govern their own AI capabilities on top of trusted, structured trade finance infrastructure, rather than attempting to replace it. This model offers a sustainable path forward for financial institutions and corporates seeking to adopt AI safely.