AI and Global trade design projects at thomson reuters

AI-powered HS suggestions

The problem
Customers who work in global trade need to classify their products with HS (Harmonized System) codes. If they do this incorrectly, they risk penalties and fines. However, there are thousands of HS codes, and it's time-consuming to comb through them. We needed AI to help reduce the time to classify, and a strong UX to streamline the process.

Design process
I worked closely with research scientists to understand how the AI would deliver its recommendations. Then, I designed a filtering system that would help increase the confidence of the AI’s suggestions, while also ensuring users feel in control. See below for a screenshot of the initial proof of concept design I delivered, which was presented at SYNERGY conference 2023.

AI HS suggestions POC

Results
The proof of concept received positive feedback, and was adapted into a side panel in the Global Trade Management product (see screenshot). It carries over many of the same features including a filtering process, confidence score, and code descriptions.

Final HS suggestion design in product

ONESOURCE Pre-Login page

The problem
ONESOURCE log in page currently just has a log in box and footer (see screenshot). The team wanted to update the site to follow our design system updates, which would increase efficiency in the long-run. We decided it was also a great opportunity to innovate, and investigate what operational and user experience needs we could support in a new user experience.

Current ONESOURCE landing page

Design process
A survey was conducted with hundreds of participants which gave significant guidance on what users prioritized. Based on the results, I emphasized features like tax news and educational resources above the fold of the page.

Results
The redesign received positive feedback, notably for being more informative and engaging than the legacy page. It was approved for development in 2026.

Screenshots available upon request.