Combination of Rule-based and Inferred Approach in Semantic Layer Creation

Semantic layer creation in AI is evolving to combine rule-based and inferred approaches. An example from a recent Snowflake summit showcased a product that leverages AI to create a semantic layer from PDFs, enabling natural language queries without predefined rules. While structured databases provide rules, new information requires manual rule creation. AI with vast data training can probabilistically determine the semantic layer. The ideal approach is a combination of rule-based and inferred methods to balance accuracy and avoid issues like AI hallucination.