Release 5.0 transforms Ask Oracle Select AI chatbot from a conversational chatbot into a complete low-code platform for building governed, visual, and agentic AI applications on Oracle AI Database and Autonomous AI Database. This release brings together visual analytics, AI profile lifecycle management, agent builders, agent orchestration, prebuilt agents, governance, application customization, and low-code development in one Oracle APEX-based experience.

Earlier releases helped you ask questions of your database data and trusted content. Release 5.0 expands that journey so you can build agentic AI applications: generate charts from natural language, configure NL2SQL and RAG without code, create agent teams, manage tools, and control which capabilities are available to different users.

What’s New in Release 5.0
Visual Agent Builder for creating Oracle Select AI agents and agent teams
Agent Team Map for understanding orchestration relationships
One-click installation and management of prebuilt Oracle AI agents
Full AI profile lifecycle management for NL2SQL, RAG, and agent experiences
Fine-grained enterprise governance and button-level access controls
Better branding, navigation, defaults, and usability for AI applications

The most important reasons to adopt Release 5.0 are:

  • Build AI agents and agent teams visually with Agent Builder, tool mapping, and team orchestration views
  • Configure NL2SQL and RAG without code through improved AI profile creation, editing, validation, and reuse
  • Deploy prebuilt Oracle Select AI agents that you can adapt to your own data, tools, and workflows
  • Govern enterprise AI applications with centralized settings, defaults, and button-level access control

The release also delivers a smoother application experience, with greater configuration control through settings, refined navigation, branding improvements, expanded default options, and stability enhancements that make the app easier to adapt for teams and solution accelerators.

Build in-database AI agents visually

The new Agent Builder is one of the biggest additions in Release 5.0. Instead of manually configuring agent metadata, developers can visually assemble agent teams manually or use natural language to generate an initial agent team.

As Oracle Select AI Agent Framework workflows become more sophisticated, you need a clear way to see how teams, agents, tasks, and tools fit together. Release 5.0 brings those workflows into the Ask Oracle Select AI APEX application so you can build, inspect, and manage agentic behavior in the same place you test it. While you can use the Oracle Select AI Agent Framework from PL/SQL and Python APIs and no-code user interfaces like Oracle AI Database Private Agent Factory, this chatbot resides directly in your database as an APEX app that you can inspect, customize, and extend to your specific needs.

Agent Builder

Agent Builder provides a low-code setup experience for creating, editing, and managing your agent teams. You can describe the agent you want in natural language, use AI assistance to generate the initial configuration, and then refine the team, agents, tasks, and tools in a guided graphical interface.

That shortens the path from idea to working agent workflow. You can assign tools, validate team behavior, and manage Oracle Select AI prebuilt agents without jumping between configuration scripts and separate orchestration tools.

Figure 1: Agent Builder interface for low-code configuration of agents, tools, and team orchestration.

Agent Team Map

The Agent Team Map gives you a graphical view of team-agent-task-tool relationships. It helps you understand how work is routed, which tools are assigned to each agent, and where each task participates in the overall orchestration flow.

Figure 2: Agent Team Map view showing the relationships among an agent team, its agents, tasks, and tools.

Prebuilt Agents Interface

The Prebuilt Agents interface helps you get started faster with Oracle Select AI prebuilt agents, which are also available in this GitHub repository. You can install them, review how they are structured, and then use Agent Builder and AI profile management to adapt them to your own data, tools, and workflows.

Figure 3: Install and manage Oracle Select AI prebuilt AI agents.

Configure NL2SQL and RAG without code

Ask Oracle Select AI supports multiple interaction styles, including NL2SQL, RAG, and agents. Release 5.0 makes profile management less about implementation details and more about productivity. You can configure AI models, switch AI providers, set up retrieval behavior, and validate profiles, whether they are part of NL2SQL, RAG, or agent interactions.

For RAG, AI profile creation and management help you configure and validate semantic similarity search-backed interactions over text-based content, so responses can be grounded in trusted sources. For NL2SQL, profile creation, update, deletion, and validation are easier to manage for natural language query generation. Agent-style experiences can use profiles together with agent team configuration, aligning model behavior, retrieval, and orchestration in one app.

Figure 4: Profile management view for NL2SQL, RAG, and agent-style experiences with clearer lifecycle actions.

Fine-grained enterprise AI governance

As AI applications move from demonstrations to shared team environments and production-style deployments, administrators need precise control over what users can see and do. Release 5.0 expands enterprise governance so you can tailor the app experience for different audiences, environments, and risk profiles.

Under UI settings, you can configure the app name, logos, and related visual identity settings, which is useful when adapting Ask Oracle Select AI for a specific team, business unit, or solution accelerator. Conversation Settings enable you to set the default conversational style, NL2SQL and RAG profiles, and agent team. Action Controls enable you to limit functionality such as code editor access or the ability to export results to PDF or Excel.

Access Control gives administrators button-level permissions for capabilities in different environments or for different audiences. A demo environment may expose exploratory features, while a production-style internal app may hide delete actions, exports, advanced diagnostics, or agent reasoning details.

Figure 5: Admin Controls view with button-level access settings for editor, exports, timer, delete, and agent reasoning visibility.

A production-ready foundation for enterprise AI applications

Release 5.0 makes the full Ask Oracle Select AI workflow to feel more connected. Home, navigation, settings, agents, and visual flows have all received stability, usability, and performance improvements. This cohesive workflow helps you move faster from question to answer, from answer to visualization, and from idea to governed AI agent experience.

The bigger story is the journey. Releases 1 through 4 helped you ask questions of Oracle AI Database data. Release 5.0 helps you build complete AI applications.

You can use Release 5.0 to:

  • Ask questions over database data using NL2SQL
  • Ground answers in trusted content with RAG
  • Run AI agents created with Oracle Select AI Agent Framework for task-oriented workflows
  • Build and manage agent teams visually
  • Govern app capabilities through centralized controls
  • Customize the experience with defaults, labels, navigation, and branding

Release 5.0 marks the evolution of Ask Oracle Select AI chatbot from a demonstration application into a production-ready foundation for enterprise AI. Whether you are building conversational analytics, RAG assistants, or multi-agent workflows, Release 5.0 provides the tools to build, govern, and deploy them from Oracle APEX on Oracle Autonomous AI Database and Oracle AI Database.

Get started

To try the latest release, import the latest Ask Oracle Select AI APEX application file into Oracle APEX in your Oracle AI Database instance. Configure your Oracle Select AI profiles, choose your default mode, set up your agent teams, review governance settings, and start asking questions.

From there, you can customize the app for your own data, workflows, branding, access controls, and AI application experience.

Resources

See the following resources: