Best Practices in AI

Martin discuss the challenges companies face in implementing AI and machine learning projects, with failure rates as high as 80%. Data issues are a major contributor to these failures, emphasizing the importance of getting the data foundation right. Building a feature store and reusing features across multiple models can improve organizational productivity and time to market. By curating valuable features and cataloging them for reuse, organizations can streamline the process of model training and deployment across different use cases. Teradata's open and connected strategy allows customers to leverage external data sources while running feature processing on Teradata. The bring your own analytics and bring your own model components enable flexibility in using tools and languages for model training and deployment. Simplifying models for inference can improve speed, performance, and reduce costs, making AI more accessible and practical for everyday use. The focus on leveraging multi-structured data can lead to meaningful insights and personalized customer experiences in various industries, including financial services. The evolution of Teradata's culture towards openness and innovation has been driven by changes in leadership and industry trends. Martin hopes to see significant progress in leveraging gen AI for customer experience at Teradata Possible next year, reflecting the company's commitment to advancing AI technologies for its customers.