Use Cases

Details
You are currently viewing the documentation for a stable release. If you are working with the main development branch, select v0.0.0.

Use cases are end-to-end demos that show the AI Optimizer working on a complete, realistic scenario. They pick up where the Walkthrough leaves off.

Once you have:

  • a chat model;
  • an embedding model, and;
  • an Oracle AI Database configured,

a use-case takes you from an ungrounded LLM to one grounded in your own structured data (NL2SQL) and unstructured documents (Vector Search).

Each use-case is self-contained; it lists the data and prompts it needs, the models it was tuned against, and example questions to ask at each step. However, you are encourage to experiment with different models, different prompts, and a variety of questions.

New here?

If you haven’t set up the AI Optimizer yet, start with the Walkthrough. It installs a local environment you can reuse.

Available Use Cases

  • Racing Championship

    Ground an LLM step by step (NL2SQL, Vector Search, then both) against a synthetic 100-driver racing championship, ending with a live championship reveal.