Using the OCI Generative AI Agents Service
Free Oracle Cloud Infrastructure 2025 Generative AI Professional practice — 6 questions on Using the OCI Generative AI Agents Service, with explanations. No sign-up. Full 12-question mixed test →
A retail company has a live Oracle Autonomous Database 23ai table that stores current inventory counts, updated every few minutes. The company wants its Generative AI Agent to answer 'How many units of SKU-4521 are in stock right now?' with the exact current count. Which agent configuration BEST meets this requirement?
The SQL tool translates natural language into SQL and executes it live against the database, returning the exact current value, which is essential for real-time structured data queries.
A company built one Generative AI Agent connected to a knowledge base of HR policy documents. They want an internal testing endpoint with content moderation disabled for QA debugging, and a separate public-facing production endpoint with the strictest moderation and citations enabled. Which approach is correct?
Agent endpoints are independently configurable for session, moderation, and citation settings while sharing the same underlying agent and knowledge base — this is the designed pattern for dev/prod separation.
An enterprise has an existing OCI Search with OpenSearch cluster that already indexes thousands of internal wiki articles using custom analyzers and synonym lists tuned by the search team. They want the Generative AI Agent to ground responses on this existing index without re-ingesting the documents. Which knowledge base data source type should they select?
The OCI Search with OpenSearch data source type lets the agent query an existing OpenSearch index directly, reusing existing analyzer and synonym tuning without re-ingesting or re-chunking the documents.
The compliance team requires that an agent endpoint block ANY user input or agent output containing hate speech, violence, or self-harm content, including borderline phrasing, before a response is returned. Which endpoint configuration correctly satisfies this?
Enabling moderation on both input and output at the strictest level filters harmful content in both directions and catches borderline phrasing, matching the compliance requirement.
A Generative AI Agent's knowledge base references a bucket in OCI Object Storage. When the agent attempts to ingest the documents, it fails with an authorization error. What is the MOST likely missing configuration?
The Agents service authenticates via a resource principal; a dynamic group matching that principal must be granted an IAM policy allowing read access to the bucket for ingestion to succeed.
A team needs a solution that maintains multi-turn conversation context across a session, calls a custom backend function mid-conversation to check order status, cites the enterprise documents it used, and enforces content moderation — all without writing custom session-management or orchestration code. Which approach is correct per exam scope?
The Agents service natively provides managed multi-turn session tracking, tool calling for custom functions, citations, and configurable content moderation at the endpoint level — meeting every requirement without custom orchestration code.
Ready for the real thing?
The full course has two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed answer explanations.
Start my full course on Udemy →