The case for using structured and semi-structured data in generative AI

The text discusses the application of generative AI on structured and semi-structured data. It explains how unstructured data is converted into numerical representations called vectors, which are then stored in vector databases for training or augmenting generative AI models. The common use case of generative AI as an answer machine is also mentioned. The text further explores the possibility of using structured and semi-structured data from SaaS applications and operational databases to power chatbots and other products that depend on knowledge about business operations. It suggests extracting and vectorizing the contents of text-rich fields from tables, or concatenating them together to build text-rich fields for this purpose. The text also highlights how structured data can be adapted to serve the needs of generative AI by removing its structure rather than imposing it. This is beneficial as most practical analysis-ready data that companies generate will remain structured in the foreseeable future. Lastly, the text introduces another approach to generative AI and structured data: using natural language to interact with numerical and categorical data for reporting purposes. It mentions how business intelligence platforms leverage generative AI to convert natural language into queries or scripts that can be used to produce charts, tables, and metrics as needed. In conclusion, the text emphasizes that generative AI offers many opportunities to relate to data in completely unprecedented ways, regardless of whether it is structured or unstructured.