Clearbit's data team has launched a significant update to its company categorization system, aiming to enhance the accuracy and efficiency of B2B lead scoring, routing, and analytics processes. Frustrated by unreliable industry data and the limitations of NAICS and SIC codes, lead Data Product Manager Ale Cabrera spearheaded the initiative to completely rebuild the system from scratch, leveraging an AI-powered classification model that accurately assigns NAICS and SIC codes to companies worldwide using multilingual descriptions. This update improves the categorization accuracy for the top 100,000 companies in Clearbit's database and is set to expand to others soon. The new system also enhances global coverage by utilizing a dataset of over 50 million companies, supporting descriptions in languages such as Spanish, Mandarin, Farsi, and Swahili. Looking forward, the team plans to refine company tags, international taxonomies, and company hierarchies, inviting users to submit specific categorization needs for further development.