To generate accurate SQL queries, LLMs require more than just raw table and column names; they need clear and concise schema context.

However, database schemas are often opaque and lack thorough documentation, forcing the model to infer the data's meaning and intent. For example, when an LLM encounters ambiguous identifiers, such as T1, C123, or emp_id, it must guess what each object means, how objects relate to each other, and what data values are valid.

AI enrichment eliminates this guesswork by allowing you to add context to your schema through human-written annotations. By providing simple, domain-specific annotations, you give the LLM the precise cues it needs to write accurate and efficient SQL queries. These annotations clarify the intent of your data, making your database schema more transparent. Thus, AI enrichment closes the gap between your schema and the context that the LLMs and AI-driven applications need.