Chris Blaisure from Elastic discusses the importance of adopting a "build and buy" strategy for AI in enterprise architecture, emphasizing the need to integrate existing SaaS investments with custom AI applications. The article highlights the challenges of fragmented AI features in SaaS applications, which often lead to low adoption and failure to deliver business impact, as evidenced by a study indicating that 95% of enterprise AI investments fail to yield returns. To address these issues, Blaisure proposes a durable enterprise architecture focused on context retrieval and performance monitoring, allowing organizations to leverage AI models with accurate business context, thereby enhancing the relevance and trustworthiness of AI outputs. By utilizing context engineering and advanced search techniques, enterprises can improve AI model accuracy, while performance monitoring through tools like Elastic Observability ensures user engagement and operational efficiency. Blaisure illustrates the success of this strategy at Elastic, where the implementation of a unified, context-aware architecture has led to significant cost savings and improved digital support delivery, showcasing the potential for scalable AI solutions that provide long-term ROI.