Success stories shaping the AI transition in trade finance
By Liza Tan
The integration of GenAI into trade finance processes can unlock new value across many areas including data integration, fraud detection and sustainability auditing. While navigating legacy systems is a significant challenge, striking the right balance is necessary for organizations to thrive.
GenAI is a multibillion-dollar opportunity for the global banking industry, and it has the potential to transform trade finance – an industry rich in structured and unstructured data, complex processes and numerous stakeholders.
The opportunity spans the entire “value chain of financial services”, according to Sandeep Mukherjee, who leads T3i’s AI practice and is also the founder of Tagor AI, a company that develops AI-native strategies, products and apps exclusively for financial services companies.
Mukherjee emphasizes that the potential value from adopting and understanding AI, especially GenAI, in financial services companies is substantial in the short, medium, and long term with the potential to capture value now given the capabilities of the current frontier models.
Read the first article in our AI series on “Banking on the future of GenAI” here.
Unlocking the potential of GenAI in trade finance
Brian Edmondson, partner at T3i, points out that a clear opportunity lies in the processing life cycle. Purchase order information can be fed into a GenAI-driven system that recognises prior transactions, enabling the creation of Letters of Credit (LCs) or guarantees, which can then be submitted to the bank for processing.
Edmondson has almost 50 years of experience in the industry and was most recently, the global head of trade and working capital finance at Finastra. He noted three other areas where GenAI has the potential to transform the trade finance industry. First, in fraud detection. Analyzing historical fraud data can help reduce fraudulent incidents. Edmondson added that there are a couple of projects which T3i is working on that taps on GenAI technology to automate checks on eligibility and exposure limits.
Second, by providing a robust audit trail for green initiatives. Edmondson cited an example involving a palm oil product where the system can verify if a supplier on a recognised green list, has been certified by an agency confirming it meets green standards. This information can also be incorporated into a repricing algorithm while maintaining an audit trail.
The third aspect is the potential to customize services for large corporations versus small and medium-sized enterprises (SMEs).
“Typically, banks have tried to come up with solutions that are one size fits all. With GenAI, you can really tailor the experience for the different types of customers and make sure that you get the best system for them,” said Edmondson.
Mukherjee highlighted Morgan Stanley’s US operations and Klarna, a ‘buy now, pay later’ service, as examples of the financial industry “going all in on AI”. Both financial services companies are utilizing AI for client engagement and operational efficiencies.
Navigating legacy systems
While many financial services companies are keen to implement emerging technologies like GenAI, there is still the significant challenge of navigating legacy systems. Mukherjee explained that many organizations face obstacles, including fragmented data across various locations, siloed departments and constrained investment capabilities. In fact, Klarna also ended its SaaS partnership with Salesforce and Workday citing its Gen-AI overhaul – likely the first wave in a series of broader disruptions to the fintech / SaaS markets.
“The beauty about the recent tech stack of AI technology is that you can access a lot of the frontier models at fairly reasonable costs using just an API call and building out the appropriate infrastructure, so it is certainly accessible to financial services companies of every size and scope,” added Mukherjee.
When it comes to enhancing interoperability to facilitate data integration with GenAI applications, especially for legacy systems lacking proper APIs, Edmondson said that many banks are investing in interoperability layers to support this and there are also several AI-based fintechs which have successfully integrated into back-office systems.
Advances in GenAI
The successes Edmondson has witnessed are mainly AI-based solutions on the compliance and document checking side. The next stage, he says, will be utilizing Large Language Models (LLMs) and using the GenAI to extend the capabilities of those incorporating more areas of compliance and to start looking into the risk side.
“The time to market has shrunk considerably using GenAI,” said Edmondson.
He explained that developing an LC application using traditional means usually takes at least three to four months. In contrast, with GenAI, this process can be completed in just a few weeks, significantly speeding up delivery based on his experience.
Key steps for successful implementation
For organizations looking to implement GenAI in their trade finance process, Mukherjee suggests three steps. First is expertise, to seek partners who understand both the technology and the industry. Second is speed, to move quickly to adapt and leverage the benefits of GenAI. Third is pragmatism, to establish a roadmap with realistic milestones and budgets to commence execution and strive for speed-to-value.
“Finding that right middle ground where they have the right use cases, a roadmap that is pragmatic and execution oriented, and calibrating it to the level of spend that they could afford to get started with. I think those are the three building blocks for our clients to get started with,” added Mukherjee.
For more insights from Sandeep and Brian, listen to the full podcast here at Trade and Treasury Now.
And if you’re interested in exploring how generative AI can transform your business, reach out to the T3i Partner Network.