Today, the AI Assistants in Oracle Analytics Cloud (OAC) and Fusion AI Data Platform (FAIDP) require local indexing of datasets or subject areas. This indexing captures synonyms, business terms and which attributes will be indexed, but it must be configured separately in each tool, leading to duplication and inconsistent AI interpretations.

Oracle AI Data Platform already provides a unified, governed catalog of metadata and lineage across the enterprise. By extending this catalog to store AI semantic profiles (for example synonyms, NLQ hints, key fields, glossary-linked definitions), organisations could optionally centralise their AI semantic governance.

This proposal introduces an optional mode where OAC and FAIDP can derive their indexing from Oracle AI Data Platform’s semantic metadata, while still preserving the existing local indexing workflow for customers who prefer standalone configuration.

• Oracle AI Data Platform is enhanced with “AI semantic profiles” attached to catalog objects.
• OAC and FAIDP can optionally pull these profiles to generate their AI indexes automatically.
• Any updates in Oracle AI Data Platform (for example new synonyms, glossary definitions or key fields) cascade to all AI Assistants for consistent behaviour.
• Local indexing remains fully supported for organisations not using Oracle AI Data Platform or requiring isolated semantic definitions.

• One place to define synonyms, business terms and NLQ behaviour for all Oracle analytics tools.
• Consistent AI interpretations across OAC, FAIDP, Oracle AI Data Platform Workbench agents and external consumers.
• Reduced duplication and maintenance effort.
• Stronger alignment between governed metadata, security and natural language experiences.
• No loss of flexibility: local indexing continues to work exactly as today.