Exploring AI in Oracle APEX 26.1

Artificial Intelligence (AI) is transforming the way organizations and developers work. Businesses and professionals are increasingly adopting AI tools to improve productivity, automate processes, and build smarter applications.

Oracle APEX introduces a significant evolution in application development by integrating governed Generative AI capabilities directly into the low-code platform. This enables developers to build intelligent, conversational applications while maintaining enterprise-grade security and control.

In this article we explore AI capabilities in Oracle APEX 26.1

What is AI Agents

AI Agents can be created to reason over user requests and take actions through approved AI Tools. Each tool exposes a specific application capability the agent is allowed to invoke, such as retrieving data, running server-side PL/SQL, or executing client-side JavaScript. APEX manages the execution flow by preparing context, dispatching tool calls, executing tools, handling results, and composing the response. This gives teams a built-in path for adding agentic behavior while keeping developers in control of the agent’s capabilities and boundaries.

AI Agent Architecture in Oracle APEX

Oracle APEX implements a structured AI execution model composed of four key layers:

• AI Agent Layer – Interprets user intent and manages conversation flow
• AI Tool Layer – Defines allowed actions and system capabilities
• Execution Layer (APEX Engine) – Executes PL/SQL, JavaScript, or API calls
• Data Layer (Oracle Database) – Processes and persists transactional data

This architecture ensures that Large Language Models (LLMs) operate strictly within predefined boundaries.

Key AI Capabilities in Oracle APEX 26.1

Oracle APEX 26.1 extends the capabilities of low-code development by enabling developers to build conversational AI agents that interact with users through natural language prompts.

AI Agents and Tools: Controlled Execution Model

AI Agents in Oracle APEX do not execute code directly. Instead, they interact with AI Tools, which expose controlled application capabilities.

 A tool may:

• Execute PL/SQL procedures

• Call REST APIs  

• Run validation before anything gets committed

• Trigger client-side interactions

The APEX engine manages the full execution lifecycle:

•         User Request

•         Agent Interpretation

•         Tool Selection

•         Execution

•         Response Assembly

This ensures all AI-driven actions remain governed, auditable, and secure

Use Cases

1. Leave Management System

Purpose: Enable an AI Agent to interact with the leave management system using natural language.

Example:
1)       “How many leave days do I have remaining?”
2)       “Apply casual leave for next Monday.”
3)       “Show my approved leave requests.”

The AI Agent can retrieve employee data, validate leave balances, and execute PL/SQL processes to create leave requests automatically.

2. Student Information System

Purpose: It supports Educational institutions can use AI Agents to improve student services.

Example:
1)       “Show my semester timetable.”
2)       “What is my current GPA?”
3)       “Show my upcoming exam timetable.”

The AI Agent can securely retrieve student information and guide users through academic processes.

Execution Flow

1. User submits natural language request
2. AI Agent interprets intent
3. Appropriate AI Tool is selected
4. APEX executes backend logic (PL/SQL/API)
5. Database processes request
6. Structured response returned
7. LLM formats final response
8. Output displayed in APEX UI

Governance and Security Model

Oracle APEX enforces strict governance across AI execution:
1)  The LLM never writes or runs SQL directly — it can only call tools you've defined
2)  Your existing APEX security context applies to every AI action, no exceptions
3)  Database privileges don't get bypassed just because a request came through an agent

4) Every AI-driven action is logged, supporting auditing and compliance requirements

 This ensures enterprise-grade security while enabling AI-driven automation.

Suggestions for Students or Beginners

Students OR beginner developers should start by learning the fundamentals of Oracle APEX before moving into AI-powered features. Understanding SQL, PL/SQL, database concepts, page development, forms, reports, and processes in Oracle APEX will make AI integration much easier.

To begin learning AI in Oracle APEX 26.1:

1) Learn the basics of Oracle Database, SQL, and PL/SQL.
2) Learn the basics of Oracle APEX.
3) Build small APEX applications such as Employee Management. 
4) Explore REST APIs and Web Services integration.
5) Understand how AI Agents and AI Tools work in Oracle APEX.
6) Practice writing prompts and testing AI interactions inside applications.
7) Start with simple AI use cases like question-answer systems.
8) Oracle APEX 26.1 provides an excellent opportunity for beginners to explore AI-powered low-code development. 

References: 

https://www.oracle.com/apex/whats-new

https://docs.oracle.com/en/database/oracle/apex/26.1

https://blogs.oracle.com/apex/coding-with-the-ai-powered-apex-assistant-on-oracle-apex