Tom Ollane, chair of the HACC pre‑event workshops, opened the session and introduced Oracle presenters who demonstrated a live build of a generative‑AI Q&A interface inside Oracle APEX.

The demonstration, led by Ashwin Rao and Santosh Kumar, both product managers for Oracle APEX, walked through configuring generative‑AI services, using retrieval‑augmented generation (RAG) to ground answers in application data and adding an AI‑powered text generator to produce school application letters. "Oracle APEX is Oracle's strategic low code application development platform," Rao said, adding that it is "available at no additional cost with Oracle Database." He summarized generative development as "using generative AI to accelerate application development." Jason Daigley, Oracle's Oahu resource, offered follow‑up support: "If you have any follow‑up questions or anything after today's workshop, don't hesitate to reach out."

Why it matters: The demo showed how low‑code tooling plus generative AI can produce interactive, data‑aware user experiences quickly. That combination can let non‑specialist teams create chatbots that answer questions based on their own records rather than relying on generic web results.

What presenters built and showed
- Workspace and AI provider configuration: Presenters opened the App Builder workspace and the new Generative AI utility. They demonstrated configuring Oracle Cloud Infrastructure (OCI) generative AI models by creating API credentials (user ID, tenancy, private key, fingerprint) from an OCI profile, and they noted the same utility accepts OpenAI and Cohere API keys for integration.

- AI‑assisted development features: Rao used the APEX assistant to generate data‑model DDL and application blueprints from natural‑language prompts and to generate SQL queries from English prompts. He demonstrated in‑editor AI help for debugging SQL.

- RAG (retrieval‑augmented generation): The team created a RAG source that runs a SQL query to fetch a concatenated single‑column context for a selected high school record. That context is sent with user prompts so the chatbot answers are grounded in the application's school data rather than generic internet text.

- Chatbot and UI integration: The session added an inline modal drawer and an embedded chat region on a faceted search results page. They created two chat entry points: an "info" icon that opens a school‑specific chat (learn mode) and a top‑right "Ask a question" button that opens a generic chat. Quick actions demonstrated included "Provide an overview of the school" and "What is the graduation rate?" In the demo a school‑specific query returned a 98% graduation‑rate example for one record; another example returned 61% for a different record (these were demonstration values shown during the session).

- Apply form and AI‑generated letter: The presenters used Quick SQL to create an applications table, built a drawer form tied to that table, and added a rich‑text editor. They configured a "Generate letter" action that sends page values (parent name/email, student name, selected school, etc.) to an AI text generation call; the returned text populates the rich‑text editor and can be submitted as the application body.

Resources and follow‑up
The presenters pointed attendees to apex.oracle.com for free workspace creation and learning materials and announced a limited promotion for free Oracle APEX certification enrollment (promo runs through Oct. 31, according to the demo). Ashwin Rao repeatedly encouraged attendees to post questions in chat during the session and use the documentation links the team provided.

Discussion vs. decisions
This event was a product demonstration and instruction session; no formal votes or policy decisions were recorded. The session combined a short presentation and a live build with audience Q&A in chat. The presenters paused for attendee questions and promised post‑session links and slides.

Ending
The session concluded after the live build and a short Q&A; organizers said they would share slides and the recording with attendees. Presenters closed by repeating the resources and contact options for follow‑up support.