This paper shares our work on building a machine learning system to categorize transactions for Intuit's QuickBooks product. Transaction categorization is challenging due to the complexity of accounting, the need for personalization, and the diversity of customers. We have broken down this monolithic problem into smaller pieces based on customers' life-cycle stages, and tailored solutions to address customer pain-points for each. Modern machine learning technologies such as deep neural networks, transfer learning, and few-shot learning are adopted to enable accurate transaction categorization. Furthermore our system learns user actions in real-time to provide relevant and in-time category recommendations. This in-session learning capability reduces user workload, improves customer experience, and helps to cultivate confidence.

In the midst of COVID-19, a global economic crisis is threatening the livelihoods of small business owners everywhere. In ordinary times, 50 percent of small businesses go out of business in the first 5 years. In today's extraordinary times, nearly 7.5 million (~25 percent) of small businesses in the U.S. alone have been at risk of closing permanently in a matter of months (Source: Main Street America's Small Business Survey 2020).