Once a company understands its challenges and goals, and has an understanding of where the data sources needed are located, bringing them together can be challenging. 

Structured data will have table and column names that differ, and they need to be represented in the same way. Different sources of data will potentially have the same customer or business record represented, and they may use different unique identifiers for their underlying data. Emails, chat transcripts, and knowledge articles are rich sources of unstructured data that are hard to search using keywords, yet they are important for generative AI applications. 

– Oct 12, 2024