- Title and Copyright Information
- Get Help
- 1 Overview
- 2 Access Requirements for AI Agent Studio
- Access Requirements for AI Agent Studio
- Provide Access to Configure AI Agents in all Products
- Provide Access to Configure AI Agents in Oracle Fusion Cloud Human Capital Management
- Provide Access to Configure AI Agents in Oracle Fusion Cloud Supply Chain & Manufacturing
- Provide Access to Configure AI Agents in Oracle Fusion Cloud Procurement
- Provide Access to Configure AI Agents in Oracle Fusion Cloud Financials
- Provide Access to Configure AI Agents in Oracle Fusion Cloud CX
- Provide Access to Configure AI Agents in Oracle Permitting and Licensing
- 3 Before You Begin
- 4 Use Cases
- 5 Expressions in AI Agent Studio
- 6 AI Agents
- 7 Use Workflow Builder
- Create Workflows
- Nodes in Workflows
- Settings in Workflows
- Configure Vector DB Writer and Reader Nodes in a Workflow
- Configure Tool Nodes in a Workflow
- Add a Policy Model to a Policy Node of Workflows
- Add a Document Schema to a Document Processing Node of Workflows
- Add a Reference Block to a Reference Node of Workflows
- Add Multi Agent Nodes in AI Agent Studio
- Configure a Return Node in a Workflow
- Configure Workflow Triggers
- Add an Approval Process Channel to a Human Approval Node of Workflows
- Add and Reuse Functions in AI Agent Studio
- 8 Use Applications Builder
- 9 Edit Agents, Workflows and Agentic Applications
- 10 Debug and Test AI Agents
- Overview of the Debugger
- Get Started with the Debugger
- Create and Use Run Profiles
- Inspect Node Results
- Debug Switch Routing
- Set and Use Breakpoints
- Rerun, Step Through, or Resume a Workflow
- Pin or Unpin a Node Output
- Edit Node Output
- Override Node Configuration
- Validate Final Chat Output
- Checklist for Publishing Workflow After Debugging
- FAQs for Debugging and Testing AI Agents
- Why is the workflow selecting the wrong branch?
- How do I troubleshoot a data or function node that returns unexpected data?
- Why is the LLM generating malformed JSON?
- How can I fix an incorrect final summary?
- How do I troubleshoot a workflow that stops at an unexpected node?
- How can I identify the cause of inconsistent downstream behavior during workflow debugging?
- How do I fix an expression or variable that resolves to the wrong value?
- What should I do if the output from one node doesn't match the input expected by the next node?
- What should I do if pinned or edited output is hiding the real problem during workflow debugging?
- How do I fix a workflow that works with a configuration override but still fails in a full run?
- How do I troubleshoot a loop or parallel branch that produces incomplete results?
- How do I fix an external service, tool, or retrieval node that fails?
- 11 Monitor and Evaluate AI Agents
- 12 Promote Published Agents
- 13 Allow External Access to Fusion Applications Agents
- 14 Collaborate with AI Agents Across Platforms Using Agent2Agent (A2A) Protocol
- 15 AI Agent Studio FAQs
- Which large language models (LLMs) are currently used or supported for AI agents?
- Which scheduled processes do I need to run to make sure the features of AI Agent Studio work as expected?
- Can I use any agent in my workflow?
- Can I migrate any workflow or supervisor agent from one environment to another?
- How do I display a welcome message in a chatbot agent?
- How do I make agents respond faster?
- How do I send an email using AI Agent Studio?
- How do I track the usage of AI Agents created in AI Agent Studio?
- How can I best reduce the risk of my agent showing hallucinations?
- How do I hide or restrict access to AI Agent Studio for specific users?
- How can I provide access to an AI Agent for all employees?
- Why can't I find the standard Employee role in the Security tab of Workflow Settings?
- Can I rename the AI chat label to something other than Ask Oracle, or change the design of the chat panel?
- What's the difference between editing a topic versus editing a system prompt?
- What file types, limits, and processing capabilities are supported in AI Agent Studio?
- Why do I see missing or misaligned structured document data in AI Agent Studio?
- Why does the external REST tool only support HTTPS transactions?
- Why is the Provider list in the LLM subtab empty?
- Why is my agent not fetching data from the document added to it?
- Why can't I see the imported workflows or supervisor agents in the target environment?
- How can I move the AI Agents link to the top of the Quick Actions list?
- Why is the Roles list in the Security tab empty?
- How can AI agents be configured or instructed to respond in a user's preferred language?
- Why is my agent not responding in my preferred language?