Acxiom, a global leader in customer intelligence and AI-enabled data-driven marketing, faced challenges while building an AI-powered identity foundation for audience segmentation. The company's Data and Identity Data Science team aimed to develop an application that could interpret natural language input from users and transform it into detailed audience segments from their expansive data catalog. However, they encountered issues with scalability, debugging, and evolving requirements, which hindered the development of a robust generative AI solution. To overcome these challenges, Acxiom adopted LangSmith, a testing & observability platform developed by LangChain, providing critical features such as observability, scalability, and flexibility. By integrating with LangSmith, Acxiom successfully transformed its complex technical vision into a scalable, user-friendly application that drives better marketing precision.