Developers and IT leaders increasingly rely on unstructured data, such as source code, README files, and code comments, to make informed decisions in software development, but this type of data is often challenging to analyze due to its lack of predefined format. Retrieval-augmented generation (RAG) is emerging as a solution, allowing for the customization of large language models (LLMs) to harness insights from unstructured data by adding context from various organizational and web sources. By utilizing RAG, developers can surface organizational best practices, accelerate understanding of codebases, and improve development and product decisions through more nuanced feedback. GitHub Copilot Enterprise, powered by RAG, exemplifies how AI tools can help developers receive natural language answers tailored to specific repositories, thus enhancing productivity and understanding of existing codebases. This approach not only aids in maintaining and modernizing legacy code but also supports efficient onboarding and resolution of technical issues, while structured data analysis remains more straightforward due to its numeric nature and established methodologies.