Cognitive Trading Using Watson

By: Maximiliano Ribeiro Aquino Santos

Trade negotiations and the complexity of rules of origin

Trade negotiations have never been known for being simple, easy, nor fast. It usually takes months or even years to close an agreement between parties. A lot of work is done behind the scenes to understand what can be achieved and reach something that is really desired and feasible for all parties involved.

Rules of origin—criteria to identify the origin (nationality) of a product—is one subject that makes trade negotiations complex. These rules have to be considered when we need to define which products are considered to be originated by which of the parties, which products can be exchanged with differentiated duties, and which restrictions can be applied to which specific products.

Let’s take a simple pen for example. Rules of origin in a trade agreement between Country A and Country B can specify that a pen is originated by Country A if the product was entirely conceived in Country A. Another rule of origin may specify that a pen be originated in Country A if a percentage of the pen’s value was aggregated in Country A (by manufacturing processes for example), even if the pen were not entirely conceived within Country A. Now, try to imagine a list of 6,000 products and the rules that may apply to them. How much time do you think it would take to read, classify, and understand a trade agreement so negotiators would be able to simply start negotiating?

Cognitive Trade Advisor

IBM Cloud Garage São Paulo has developed a solution that uses Watson Natural Language Processing capabilities to understand and classify rules of origin in a couple of seconds—the Cognitive Trade Advisor.

Watson helps trade negotiators by extracting and classifying rules of origin, correlating them with products. It also presents a chart so our users can understand the profiles and interests of negotiating parties. The goal is to not only be faster but also create better agreements for all parties involved.

The diagram below explains an overview of the developed solution:

Cognitive Trade Advisor

Utilizing microservices in the cognitive trade platform

Considering that the solution aimed to be the first feature of a cognitive trade platform, microservices seemed to be a great fit. Other features can be built and deployed, maximizing platform evolution capabilities. Throughout the development of the project, two negotiators were responsible for using Watson Knowledge Studio to annotate rules of origin from trade agreements and related products. These negotiators also classified rules of origin according to the most recently available literature. The agreements were annotated in both English and Spanish.

Negotiators can access this model using a front-end built on Angular 6 and deployed to IBM Cloud’s Cloud Foundry Environment, where they can upload new trade agreements that were analyzed by Watson’s NLU and NLC engines. These can then be saved into a Cloudant database. Negotiators can also use the same front-end to check trade agreements, rules of origin, and the negotiating profiles for the party of interest.

We at the IBM Cloud Garage São Paulo are very proud of having the opportunity of being the first team that applied cognitive technologies to help the development of International Commerce!

If you would like to see how the IBM Cloud Garage can help you, schedule a free, four-hour consultationhere.

Be the first to hear about news, product updates, and innovation from IBM Cloud