Putting ChatPwC to the test

Picture this:

You and your team are Customs and Trade specialists at PwC in charge of classifying your clients’ new products when they enter the market. Your task is to organize and categorize the different products—each with their own unique features and specifications—traded between countries. However, the process of classifying tariff codes is a lot like solving a puzzle with 10,000 pieces: it takes time (and requires a lot of patience).

Tax Consulting director Jonathan Wiens, Tax Consulting manager Kathleen Lawler, and Products & Technology director David Resseguie, are the “puzzle masters” in this scenario. With their extensive experience, they know to consult the Harmonized Tariff Schedule of the United States (HTSUS) as a required resource to accurately classify a range of their clients’ products.

With 97 chapters in total, the HTSUS is a comprehensive resource based on the international Harmonized System (HS) that helps customs officials know how much tax or duty should be paid on each product. It also helps businesses understand the rules and regulations for trading different products. Understanding and correctly applying tariff code classification is crucial, and—when done incorrectly—can have significant consequences.

“Classification not only impacts a company's bottom line by affecting the duties and taxes they pay, but it also ensures compliance with international trade regulation,” explains Jonathan.

Eager to try out the capabilities of GenAI in their work, Jonathan, Kathleen and David wanted to experiment with tariff code classification using ChatPwC. Would GenAI be able to correctly assign tariff codes from the HTSUS to potential products in a proof of concept?

Enter ChatPwC:

Jonathan, Kathleen and David uploaded a single chapter of the 97 chapters found within HTSUS. Putting their Prompting 101 skills to the test, they asked ChatPwC to help them summarize the fundamental information about the products and suggest the appropriate tariff codes associated with each.

The conversation went something like this*:

Customs and Trade team: Please complete a tariff code classification for the products below, analyzing their descriptions, features and specifications in the context of HTSUS chapter 4.

ChatPwC: Based on the descriptions, features and specifications, I suggest that Product A falls under tariff code 12345 and Product B falls under tariff code 67890.

Customs and Trade team: That's a good starting point. We reviewed the suggested tariff codes to make sure they align with the rules and regulations in HTSUS chapter 10. Product A looks good. However, we think Product B aligns better with tariff code 54321, due to criteria from another chapter of the HTSUS not uploaded here.

David, Kathleen and Jonathan continued to input other product information into ChatPwC to further test the tool’s capabilities.

ChatPwC helped them summarize the products’ attributes and specs from the data and made recommendations on how they should be classified. This allowed David, Kathleen and Jonathan to concentrate their efforts on reviewing and analyzing the accuracy of the tariff code classification that ChatPwC provided.

The impact:

"ChatPwC didn't replace our process of judgment-driven work; however, it did significantly streamline the analysis upfront,” said David. The tool not only saved them time in the complex and often lengthy process of tariff code classification, but it also showed them how they could use GenAI to approach this “puzzle” more strategically.

David, Kathleen and Jonathan’s experience with ChatPwC is just one example of how PwCers are exploring new ways to use GenAI in their day-to-day work. It’s also a prime example of our human-led and tech-powered approach—when used responsibly and prompted correctly, GenAI technology has the potential to enhance human productivity and accuracy.

*These tariff codes and chapter numbers are for illustrative purposes only.