A successful Oracle Analytics Cloud (OAC) AI Agent is built upon three essential components that function as the actual "product," while the agent itself is merely the container. These three components are:
- The Dataset: This is the foundation of the agent and must be designed as a governed, flat view where every field uses business language rather than technical jargon. It is critical that the data is self-describing, meaning categorical values like "Y/N" are replaced with clear terms like "Paid On Time" or "Paid Late," and all complex KPIs are pre-calculated so the agent does not have to guess business formulas.
- The Knowledge Document: This component serves as a retrieval index rather than a standard narrative document. To be effective for Retrieval-Augmented Generation (RAG), it must be structured with short, direct, and independent "chunks" of information—such as specific policy rules, business definitions, and escalation procedures—that the agent can easily retrieve to answer questions.
- Supplemental Instructions: These instructions act as the bridge between how business users naturally speak and how the data is technically structured. They are used to define KPI formulas, compliance thresholds, and behavioural rules, as well as to map user vocabulary to specific data fields and values.
The sources emphasize that the design of these three components is fundamentally a Business Analyst's responsibility because it requires deep domain knowledge rather than just technical configuration.
Tags: Oracle Analytics Cloud | OAC AI Agent | Business Intelligence | Oracle EBS | OCI GenAI | Business Analyst | Dataset Design | Knowledge Management | RAG | Supplemental Instructions
