Select AI Agent Use Cases

Select AI Agent supports practical scenarios across roles: connect external services, build task-focused automation with PL/SQL tools, and enable conversational access to data using natural-language interactions.

Use cases include:

• Integrate external APIs for enriched responses:

  Connect REST services such as shipping status, knowledge articles, or messaging. The agent blends external results with database facts and returns timely, actionable guidance.

• Build task-specific agents using PL/SQL tools:

  Create focused agents that call PL/SQL tools for updates, validations, or transformations, keeping logic close to data, improving performance, and simplifying maintenance across environments consistently. Wrap existing PL/SQL for updates and checks. Agents coordinate tasks, enforce parameters, and record outcomes while staying within transactional controls.

• Enable conversational access to enterprise data:

  Expose data through natural-language questions grounded on database objects. Use NL2SQL for retrieval, apply policies, and present findings clearly while respecting security and governance requirements. NL2SQL translates questions to SQL on approved objects. Agents summarize results and next steps while following object lists and access rules.

• Tool discovery and tool Calls

  Use tool discovery and tool call APIs when your application needs to discover available Select AI Agent tools or invoke a specific tool from SQL, PL/SQL, or an integration layer. Specify the tool name and provide input in the JSON format expected by the tool definition. If no tools have been created, the APIs return an empty result.

• Agent team discovery and state lookup

  Use agent team discovery APIs to list available Select AI agent teams, retrieve information about a specific team, or check the latest run state for a conversation. Team descriptions include the skills exposed by the team. These skills represent the underlying Select AI Agent tools and are not separate database objects.

• Agent-direct tool calls

  Use agent-direct tool calls when your application requires an open-ended conversational session and determines when the interaction ends. Configure the team entry without a task to enable this mode. Use task-based workflows when the conversation should follow predefined task instructions and completion criteria.

• Memory depth configuration

  Configure conversation memory to control how much previous conversation context is available to the LLM. Agent settings control short-term memory, while team settings control long-term memory. By default, Select AI Agent retains the most recent 30 conversation turns. Use the default value for most workloads. Increase the memory depth for longer conversational interactions that require additional context, or decrease it for shorter, task-oriented conversations.