Key Takeaways
- Centalion (formerly Gunvor) deployed an AI‑agent on the Komgo trade‑finance platform to automate the issuance and amendment of bank guarantees and standby letters of credit.
- On the first facility covered, processing time dropped roughly 55 % for new issuances and 80 % for amendments, according to the company’s internal metrics.
- The agent lives in Centalion’s own cloud, pairs a large‑language model with the firm’s treasury workflows and facility rules, and connects to Komgo via the platform’s Model Context Protocol (MCP) connector—eliminating the need for a custom integration.
- Treasurers typically spend 60‑70 % of their time on manual, low‑value tasks; the AI agent aims to free more than 60 working days per year for higher‑value risk oversight and financing strategy.
- Unlike a simple chatbot, the agent prepares transactions while final approval and accountability remain with human staff, reducing operational risk by eliminating manual re‑keying of data.
Issuing a Bank Guarantee Remains a Desk Job
At a large commodity trader, issuing a bank guarantee is still a repetitive desk job. A treasury employee first checks how much room is left on a credit facility, then pulls banking details from one system, confirms limits in another, and finally keys the transaction into a third system. At Centalion—the Geneva‑based trader formerly known as Gunvor, which booked $144 billion in revenue in 2025—this routine repeats for every instrument, creating a steady stream of low‑value work that consumes significant staff time.
Centalion Hands the Legwork to an AI Agent
On Oct. 14, Centalion announced it had transferred that legwork to an artificial‑intelligence agent. “The agent runs on Komgo, the trade finance platform Centalion has used for more than five years,” the company said, “and lets treasury staff ask in plain language whether capacity exists to issue or amend an instrument.” The agent then pulls the relevant data, checks facility rules, and prepares the transaction for a human to approve.
Early Gains Show Significant Time Savings
On the first facility covered, Centalion reported that issuance processing time fell about 55 % and amendment processing time fell about 80 %. The firm cautioned that these figures are its own and have not been independently verified, but they illustrate the potential impact of automating repetitive data‑gathering and entry steps.
Digitized Workflows Still Left Staff Jumping Between Systems
Even before the AI agent, Centalion had digitized its guarantees and standby letters of credit on Komgo’s platform. The problem, however, was that staff still had to jump between systems to retrieve data, check it, and chase status updates. As the company put it, “Digitization was done. Working the tools was the new cost.” The manual switchover between applications remained a bottleneck despite the underlying digitization.
How the Agent Integrates with Existing Infrastructure
The AI agent lives in Centalion’s own cloud environment. It pairs a large‑language model with the company’s treasury workflows and facility rule set. Crucially, it connects to Komgo through the platform’s Model Context Protocol (MCP) connector, which allows outside AI tools to read from and write to Komgo. Because of this connector, Centalion avoided a costly custom integration project. Komgo, founded in 2018 by a consortium of banks and commodity traders, now serves more than 400 clients, according to Centalion’s announcement.
Leadership Vision: Removing Friction Without Removing Accountability
Olivier Thevenon, Centalion’s treasury manager, described the initiative’s goal succinctly: “Removing friction without removing accountability.” He emphasized that while the agent handles data gathering and preparation, final decisions remain with human treasury staff. Once the agent covers all of Centalion’s trade‑finance activity, the company projects it will free more than 60 working days of treasury capacity a year for higher‑value activities such as risk oversight and financing strategy.
Treasurers Spend Much of Their Time on Low‑Value Tasks
Centalion’s starting point reflects a broader industry pattern. Treasury functions typically devote 60% to 70% of their time to manual, low‑value activities, according to EY India’s “An Agentic AI Adoption Playbook for CFOs and Treasurers.” Furthermore, 63% of finance leaders polled at J.P. Morgan’s EMEA Treasurers Forum expect employee productivity to be the area technology will change most in the next 12 months, as reported by PYMNTS. AI agents are beginning to take on core treasury duties—such as intraday liquidity decisions and foreign‑exchange forecasting—inside large banks and corporate treasury teams across Europe.
Why This Is More Than a Chatbot
What distinguishes Centalion’s solution from a simple chatbot is where the controls reside. The company stressed that checks and approvals are built into the workflow itself rather than left to the model. The agent prepares the transaction; a person decides whether to approve it. By eliminating manual re‑keying, the agent also lowers operational risk, since data moves directly from the Komgo platform to the transaction without anyone retyping it.
Scaling the Agent Across Trade Finance
Centalion labels its initial deployment “the first production use case, not the end state.” Thevenon was clear about what the firm wanted to avoid: “We didn’t want to just build another chatbot.” The plan is to extend the AI agent across the remainder of Centalion’s trade‑finance portfolio, automating similar processes for other instruments and facilities. As the agent scales, the expected time savings and risk reductions should compound, allowing treasury professionals to focus more on strategic analysis and less on routine data handling.
Conclusion
Centalion’s experiment illustrates how a well‑designed AI agent—integrated via an existing platform’s MCP connector—can cut processing times dramatically while preserving human oversight. By removing the friction of toggling between multiple systems and eliminating repetitive data entry, the trader aims to reclaim valuable treasury capacity for higher‑impact work. If the early results hold across the full trade‑finance suite, the initiative could serve as a template for other commodity traders and corporate treasurers seeking to boost efficiency without sacrificing control.
https://www.pymnts.com/news/artificial-intelligence/2026/ai-agent-cuts-centalions-trade-finance-amendment-time-80/