Press Release

AI / Machine Learning Streamlines Trade Finance

Trade finance is one of the most complex procedures in enterprise finance. According to Maersk, 200 documents are exchanged regularly among supply chain entities. These include payments, letters of credit, foreign exchange, board of compliance documents, shipping documents, and import and export documents. They must be approved by multiple entities including suppliers, manufacturers, distributors, customers, and government agencies. Processes such as determining credit for a potential trading partner are complex and time consuming. And these processes are almost entirely manual.

At best, this is expensive, time consuming, and frustrating. It adds complexity to every step of the trading process, opens the door to potentially costly errors, distracts staff from more valuable tasks. There must be a better way.

Applying digital technology

In fact, there is. Artificial intelligence (AI) and machine learning (ML) and their analogs, natural language processing and chatbots, promise to revolutionize business. They are good at exactly the processes where humans are poor – complicated, well defined tasks that demand constant attention to detail. Humans tend to get bored, minds wander, and errors are made. AI can do them much faster than humans and never miss a beat. These systems can make fairly complex decisions based on well-defined parameters (such as recommending a potential trading partner’s credit limit based on data automatically gathered from multiple trusted sources).

With natural language interfaces, it can communicate recommendations directly to the manager responsible for the final decision. You can ask ad hoc questions and get answers based on the best available data. Predictive analytics can help you anticipate and plan for specific problems and opportunities in the market rather than constantly reacting after the fact. AI can help you understand your trading partner’s concerns and probable actions, recognize patterns in orders, and better anticipate demand.

AI can automate completing all those trading documents and make sure that the right electronic forms are received by the right entities at the right time in the trading process, without missing anyone. At any time it can produce a detailed report showing exactly where each active trade is in terms of what is filed, what has been processed, what still needs to be done. And this all happens literally at the speed of light, cutting the time required to the fraction of what it is today.

Natural language processing (NLP) can understand both spoken conversation and printed documents, translate them fill them out, and send them to the correct people in the supply chain in the language of the recipient. AI and block chain can be used to increase transaction security using smart contracts that have can use embedded computer intelligence to strengthen the trust on which trading is based.

Digital Assistants for Business

New digital assistants for business like Emagia’s Gia, which are just entering the market, are the equivalent of Alexa with business skills. They become the access point for AI tools and can perform multiple functions including credit and business license verification, handling passwords in multiple languages, and other tasks associated with both domestic and international trade finance.

For example, Emagia is working with a Napa Valley winery that ships wines worldwide. It has a manual onboarding process for new distributors and retailers that includes a detailed credit check process to determine the level of credit the winery should extend to the new partner. This is very time consuming, delaying the opening of new business channels and creating other business problems.

Gia can capture all the data required for the credit application from multiple sources such as Experian and other credit rating entities. It can confirm that the partner has the required business, alcohol consumption, and any other licenses. And it does all of this in minutes rather than days. It delivers its recommendations with full drill-down capabilities into the data in natural language, allowing business executives to work directly with it rather than needing to work through a programmer. It can communicate with the prospective trading partner in her native language, simplifying and speeding the process while minimizing misunderstanding.

Conclusions: The AI/human partnership

AI and MI are valuable new tools in business that can do amazing things. However, they are not good at many things where humans excel. Most important, they cannot handle the “big picture” questions that look beyond specific processes and give them their value.

Gia, for instance, can automate the process of on-boarding new trading partners, solving a major business problem for many companies. But it cannot answer the question of how that new trading partner fits into the company’s business strategy. Does the company want to expand business in the geography this partner covers? Can this new partner replace another trading partner that is proving problematic? Can the company support the added business this alliance will bring? Will adding this partner require the business to expand production, and is that something it wants to invest in? Will some other opportunity provide a better return?

The biggest advantage of AI/ML and NLP systems is that they can free your employees from the constant distraction of completing boring, repetitive tasks and processes to focus on these and other higher level concerns and activities. Thus they allow you to increase the value of your employees while improving efficiency in many routine business processes that the company depends on.