IDUG NA 2025

Relational databases serve as the primary data source for enterprise AI projects. Traditionally, when an AI project requires data from a relational database, engineers duplicate the data onto a separate system for model development and inference. However, this duplication process presents various challenges, including network latency, complex data governance, and the need for multiple APIs. These challenges often impede the successful implementation of AI models in production for many enterprise projects. To address these obstacles, IBM Db2 offers in–database AI capabilities and external tools. With Db2, ML practitioners can create and deploy AI models directly within their databases, significantly simplifying the complexities of AI pipelines. By leveraging in–db AI features, enterprises can safeguard sensitive data, expedite inferencing processes, and seamlessly integrate AI models with downstream business applications. In this talk, I will present the AI capabilities and tools provided by Db2. I will showcase how these capabilities can be effectively utilized in upcoming AI projects, demonstrating their potential to transform enterprise AI implementations.