Getting Started with Oracle AI Agent Studio: Tools, Templates, Deployment Options and Security Setup

Oracle AI Agent Studio is designed to help organizations create, configure and manage AI agents inside Oracle Fusion Applications using the same tools Oracle uses. The material positions AI Agent Studio as a built-in capability for creating enterprise AI agents with explainability, human-in-the-loop support, security, trust, LLMs, integrations, APIs, testing, validation, prebuilt templates, extensibility, credential store, tools, topics, prompts, knowledge store and monitoring. 

For businesses adopting Oracle Fusion AI, this is important because AI agents are no longer isolated experiments. They are becoming part of daily enterprise workflows, helping users access business data, retrieve documents, trigger actions, send emails, navigate to Fusion pages, connect with external REST services and perform calculations. Oracle’s AI Agent Studio provides the foundation to build these agents in a governed, secure and application-aware way. 

What Is Oracle AI Agent Studio? 

Oracle AI Agent Studio allows organizations to create AI agents using a structured framework of tools, templates, prompts, topics, security settings and runtime capabilities. The platform is built for Fusion Applications and supports both prebuilt and custom agent experiences. The session agenda focuses on four core areas: an introduction to AI Agent Studio, a walkthrough of AI Agent Studio, guidance on when to use each tool and security configuration. 

This structure is useful because successful AI adoption requires more than creating a chatbot. Businesses need to understand: 

• What agents are available 
• How to use pre-delivered templates 
• Which tool fits which use case 
• How agents can be deployed 
• How access and security should be configured 
• How external API authorization works 

AI Agent Studio brings these elements together into one enterprise-ready platform. 

Centralized Access to AI Agents 

Oracle provides users with centralized access to all AI agents available to them based on their role. Users can access AI agents through Me > AI Agent Studio and select the Explore tab to view agents available for the logged-in user. This helps users find the right agent for their purpose without searching across different areas of the application. This is important for user adoption. If users cannot easily find the right agent, AI adoption remains limited. A centralized Explore experience makes AI agents more discoverable and easier to use. For business teams, this means employees can access AI support in a role-based way. They only see what they are permitted to access, helping keep the experience relevant and governed. 

Pre-Delivered Agent Templates 

Oracle AI Agent Studio also provides access to pre-delivered agent templates through Home > Tools > AI Agent Studio. These templates allow organizations to quickly leverage ready-made agents, deploy agents in days with minimal effort and benefit from native integration with business data, application context and security. This is one of the strongest adoption accelerators. Instead of building every agent from scratch, businesses can start with prebuilt templates and adapt them based on process needs. This reduces time to value and helps teams learn how Oracle structures agent capabilities. Pre-delivered templates are especially useful for organizations that want to pilot AI quickly while still working within Oracle’s application security and context model. 

AI Agent Studio Tools 

Oracle AI Agent Studio includes several tools that provide agents with capabilities across Fusion Applications and beyond. The material lists the following tools: Document Tool, Business Object Tool, Email Tool, Deep Link Tool, External REST Tool, User Session Tool and Calculator Tool. 

Each tool serves a different purpose. 

• The Document Tool uploads or references unstructured documents for semantic search and retrieval. This helps ground answers in approved content. 
• The Business Object Tool retrieves data from the Fusion database and performs transactions on Fusion business objects defined in the Fusion OpenAPI specification. 
• The Email Tool enables an agent to write and send emails to human recipients. 
• The Deep Link Tool sends users to a Fusion form interface to perform specific actions with the required context. 
• The External REST Tool connects to any service with a public REST interface. 
• The User Session Tool gives agents context about the end user within Fusion. 
• The Calculator Tool performs calculations for accurate math-based answers. 

Together, these tools allow agents to move beyond simple Q&A and support real business workflows. 

When to Use Each AI Agent Studio Tool 

Oracle’s material provides a clear comparison of when each tool should be used. 

• The Business Object Tool is used when agents need to interact with core business data and operations. It supports CRUD operations such as GET, POST, PUT and PATCH, complex queries with multiple parameters, fixed queries and dynamic parameters. Common use cases include retrieving employee or user details, fetching account information, managing payroll data and accessing worker assignments. 
• The Deep Link Tool is used to generate navigation links to specific application pages. It supports parameterized URLs, context-aware navigation and user-specific parameters. It is useful for directing users to payslips, account management pages or specific workflows. 
• The Email Tool handles automated email communication. It supports system notifications, predefined alert codes and templated messages. Use cases include payroll notifications, account change alerts and representative notifications. 
• The Document/RAG Tool manages document-based operations and queries. It searches document repositories, answers questions using document content and supports retrieval-augmented generation. It is useful for policy documents, FAQs and knowledge base articles. 
• The External REST Tool integrates with external systems and services. It makes HTTP requests to external endpoints, handles authentication and headers and supports multiple HTTP methods. Common use cases include third-party integrations, external APIs and cross-system data retrieval. 
• The Calculator Tool performs numerical computations and business calculations, such as leave balances, tax amounts and financial figures. 

This tool selection guidance is important because every agent should be designed with the right capability model. Using the wrong tool can create complexity, security gaps or poor user experience. 

Deployment Options for AI Agents 

Oracle’s material identifies multiple deployment options for interacting with AI agents. Users can access agents through Me > AI Agent Studio, Guided Journeys, embedded application experiences, Smart Actions in CX and externally through REST API. This is significant because AI agents need to meet users where they work. The material shows agents accessed directly through the Explore tab, from Guided Journeys, within transactions and through Smart Actions in CX. The page visuals demonstrate how agents can appear inside business workflows rather than requiring users to leave the application context. The roadmap-style platform view also shows external REST API usage, webhook triggers, social media or news alerts, app events such as late payment, low satisfaction score or delayed shipment, and direct third-party agent interaction secured through Oracle application security via IDCS. This shows the broader direction: AI agents can be accessed through application UI, workflows, external triggers and platform integrations. 

Security Configuration: Why It Matters 

AI Agent Studio security configuration is a critical part of enterprise adoption. The material highlights a common issue where an agent is built, but the assistant is unavailable because the required access has not been configured correctly. The goal is to give the access needed so the agent can run properly. This is a practical point. AI agents may fail not because the agent logic is wrong, but because the user, admin role, runtime duty role, permission group or external authorization privilege is not configured correctly. Oracle’s security setup guidance covers post-25C prerequisites, pillar-specific admin roles, an admin role for all pillars, exploring agents, TRAP for external API authorization and the overall summary. 

Post-25C Upgrade Prerequisites 

Oracle identifies key prerequisites after the 25C upgrade. 

• First, organizations must set the profile option ORA_ASE_SAS_INTEGRATION_ENABLED to Yes at the Site profile level. This prepares the Security Console to work with permission groups and related objects. The steps include going to Setup and Maintenance, searching for Manage Administrator Profile Values, entering the profile option code and saving the Site-level value as Yes. 
• Second, Oracle instructs users to run two batch jobs in sequence: Import Resource Application Security Data and Import User and Role Application Security Data. The first imports resources from LDAP into application security tables. The second copies users, roles, privileges and data security policies from LDAP, policy store and Applications Core Grants schema into Oracle Fusion Applications Security tables. 

These prerequisites are important because AI Agent Studio depends on accurate and available security data. Without these steps, role search and access configuration may not work reliably. 

Pillar-Specific Admin Roles 

Oracle provides pillar-specific admin role guidance for Human Capital Management, Supply Chain Management and Procurement. 

• For HCM, organizations can create a custom role to grant AI Studio Admin access for HCM and Common Objects while restricting access for other pillars. 
• For users with the HCM Application Administrator role, the custom role should connect to the Fai GenAI Agent HCM Admin Duty. For users without that application administrator role, the custom role should connect to both the Fai GenAI Agent HCM Admin Duty and the HCM Intelligent Agent Management Duty. 
• For SCM, users with the Supply Chain Application Administrator role should edit the Supply Chain Application Administrator job role and enable permission groups. Users without the Supply Chain Application Administrator role should create a custom role, enable permission groups and connect SCM Intelligent Agent Management duties plus the Fai GenAI Agent SCM Administrator Duty. 
• For Procurement, users with the Procurement Application Administrator role should enable permission groups on the Procurement Application Administrator job role. Users without that role should create a custom role and connect PRC Intelligent Agent Management duties plus the Fai GenAI Agent PRC Administrator Duty. 

This role structure supports controlled access by pillar so administrators can manage only the AI Studio capabilities relevant to their domain. 

Admin Role for All Pillars 

Oracle also provides guidance for an AI Agent Studio Administrator role across all pillars. Users with the Application Implementation Consultant role can create a custom role, enable permission groups and connect administrator duties for CX, FIN, GRC, HCM, PRC, PRJ, PSC and SCM. Users without the Application Implementation Consultant role should also connect Manage All Intelligent Agents under Roles and Privileges, plus the same administrator duties under Roles and Permission Groups. This is useful for enterprise AI teams or platform administrators who need cross-pillar visibility and governance. However, this kind of access should be controlled carefully because it spans multiple functional areas. 

Creating an AI Agent Explorer Role 

Not every user needs admin access. End users usually need explorer access to interact with available agents. Oracle notes that basic access for employees or other end users requires the profile option ORA_HCM_VBCS_PWA_ENABLED to be set to Y. To create an AI Agent Explorer role, organizations should create a custom role, enable permission groups, add the function security policy Access Intelligent Agent Chat with privilege code HRC_ACCESS_AI_AGENT_CHAT_PRIV, and connect the Fai GenAI Agent Runtime Duty under Roles and Permission Groups. 

After assigning the custom role to the end user or employee, users may need to clear browser cache and wait briefly before checking Me > AI Agent Studio. Oracle also clarifies that the Explorer role does not let users see all deployed agents. Users will only see the agents they are allowed to access, usually managed at the agent team level. This is an important governance point. Explorer access gives entry into the AI agent experience, but actual agent visibility remains controlled. 

TRAP: Token Relay and Proxy for External REST 

For external API authorization, Oracle introduces TRAP, or Token Relay and Proxy. TRAP allows customers to extend Fusion AI Agent Studio to invoke third-party REST services. These REST endpoints are expected to have protected access, and TRAP handles the calls with secure access. This matters for agents that use the External REST Tool. Many enterprise agents need to connect with third-party services, external systems, APIs or data sources. Since those endpoints are protected, authorization must be handled securely rather than through informal credential sharing. Oracle also states that access to authorization details is secured through the Create and Edit Backends for Visual Builder Studio privilege, with privilege code ORA_FND_TRAP_PRIV. This privilege can be added to a custom AI Administrator role or managed through a separate role for more granular control. This gives organizations a way to separate general AI administration from external REST authorization control. 

Summary of Security Role Requirements 

Oracle’s summary table brings together admin and explorer role requirements across HCM, SCM, PRC, all pillars and explorer users. A key note states that if Agent Explorer capabilities are included within an Admin role, the runtime or explorer duty roles from the final row of the table must also be assigned to the Admin role. 

The summary also confirms key mappings: 

• HCM admin access depends on HCM administrator status and the relevant HCM agent duties. 
• SCM admin access depends on SCM administrator status and SCM intelligent agent duties. 
• PRC admin access depends on Procurement administrator status and PRC intelligent agent duties. 
• All-pillar access uses administrator duties across CX, FIN, GRC, HCM, PRC, PRJ, PSC and SCM. 
• Explorer users need HRC_ACCESS_AI_AGENT_CHAT_PRIV and ORA_DR_FAI_GENERATIVE_AI_AGENT_RUNTIME_DUTY. 
• TRAP authorization access uses ORA_FND_TRAP_PRIV. 

This summary is especially useful for implementation teams because it consolidates role design into one reference. 

AI Agent Studio Resources 

Oracle’s resource page highlights documentation, training and Q&A resources for AI Agent Studio. Documentation resources include the AI Resource page in Cloud Customer Connect, What’s New for 25C, Oracle AI for Fusion Applications Get Started, guidance on using AI Agent Studio, and resources related to AI innovation, governance and Fusion AI security principles. Training resources include Oracle Fusion AI Agent Studio Foundations Associate, developer training and certification, AI Agent Studio previews and deep-dive training. Q&A resources include the Oracle AI for Fusion Apps forum, FAQs and AI Agent Studio events or office hours. These resources are important because AI Agent Studio adoption requires functional, technical and security understanding. Teams need training not only on how to build agents, but also how to govern, deploy and secure them. 

Why AI Agent Studio Matters for Oracle Fusion Customers 

Oracle AI Agent Studio matters because it gives businesses a structured way to move from AI curiosity to controlled AI deployment. 

The platform supports: 

• Discovering and using role-based AI agents 
• Deploying pre-delivered templates 
• Creating custom agents 
• Selecting the right tool for each use case 
• Embedding agents inside Fusion workflows 
• Connecting to external REST APIs 
• Managing security by role and pillar 
• Controlling runtime access 
• Supporting governed external authorization 

This combination is essential for enterprise AI. Businesses need AI that works within business processes, respects application security, uses trusted tools and supports real operational outcomes. 

How NexInfo Helps with Oracle AI Agent Studio 

NexInfo helps organizations assess, plan and implement Oracle AI Agent Studio capabilities across Fusion Applications. 

For AI Agent Studio adoption, NexInfo can support: 

• AI Agent Studio readiness assessment 
• Agent uses case discovery 
• Pre-delivered template evaluation 
• Custom agent design 
• Tool selection and configuration 
• Business Object Tool and External REST Tool planning 
• Security role design for HCM, SCM, PRC and all-pillar access 
• Explorer role setup and runtime access planning 
• TRAP authorization planning for external REST services 
• Testing, validation and user training 
• Post-go-live support and managed services 

NexInfo helps businesses move beyond AI experimentation by aligning AI Agent Studio capabilities with governance, security, business processes and adoption strategy. 

Conclusion 

Oracle AI Agent Studio provides a built-in foundation for creating, deploying and governing AI agents across Oracle Fusion Applications. The material shows how users can access agents, use pre-delivered templates, select the right tool for the right job, deploy agents across application experiences and configure security roles properly. It also highlights the importance of TRAP for secure external REST authorization. For Oracle customers, the biggest takeaway is that AI Agent Studio adoption should be approached as a structured enterprise program. Success depends on the right use cases, the right tools, the right security configuration, the right deployment model and the right support strategy. With proper planning, Oracle AI Agent Studio can help organizations bring AI into everyday Fusion workflows with stronger control, faster adoption and better business alignment. 

FAQs

What is Oracle AI Agent Studio?

Oracle AI Agent Studio is a platform for creating, configuring and managing AI agents in Oracle Fusion Applications using tools, templates, prompts, topics, security and monitoring capabilities.

How do users access AI agents?

Users can access available agents through Me > AI Agent Studio and use the Explore tab to see agents available to them based on their role.

What are pre-delivered AI agent templates?

Pre-delivered templates are ready-made agent starting points that help organizations deploy agents faster with native business data, application context and security.

What tools are available in AI Agent Studio?

AI Agent Studio includes Document Tool, Business Object Tool, Email Tool, Deep Link Tool, External REST Tool, User Session Tool and Calculator Tool.

When should the Business Object Tool be used?

The Business Object Tool should be used when an agent needs to retrieve or transact on Fusion business data using Fusion business objects.

When should the External REST Tool be used?

The External REST Tool should be used when an agent needs to connect with external systems or third-party REST services.

What are the deployment options for AI agents?

AI agents can be accessed through Me > AI Agent Studio, Guided Journeys, embedded application experiences, CX Smart Actions and externally through REST API.

What is the AI Agent Explorer role?

The AI Agent Explorer role gives end users access to interact with AI agents. It includes the Access Intelligent Agent Chat privilege and the Fai GenAI Agent Runtime Duty.

What is TRAP in AI Agent Studio?

TRAP stands for Token Relay and Proxy. It supports secure authorization when AI Agent Studio invokes protected third-party REST services.

How can NexInfo help with Oracle AI Agent Studio?

NexInfo can help with readiness assessment, use case discovery, tool configuration, security role setup, TRAP planning, testing, training and managed services for Oracle AI Agent Studio adoption.