Oracle 1Z0-1155-2 Practice Test: AI Agent Studio

1Z0-1155-2 is Oracle’s current Oracle Fusion AI Agent Studio Developer Professional — Rel 26-2 exam. Oracle is the issuer; Oracle Fusion Applications AI is the product-family lane used on this site so exam-code pages stay grouped by the application stack candidates actually study.

Use the 12 original sample questions below to preview developer decisions across AI agent scope, grounding, tools, integrations, security, testing, deployment, and operational review. Use the Notify me form if this is your target Fusion AI route.

Exam Snapshot

What To Review First

Sample Exam Questions

Try these 12 original sample questions for Oracle 1Z0-1155-2. They are designed for self-assessment and are not official exam questions.

Question 1

Topic: Agent scope

A Fusion Applications team wants an AI agent to answer invoice-status questions and open a workflow only when a user has the required role. What should the developer define first?

• A. A shared administrator account for every user
• B. A random answer style so the agent feels conversational
• C. The agent scope, data sources, allowed actions, identity boundary, and escalation points
• D. A spreadsheet export of all invoices for the prompt

Best answer: C

Explanation: A developer professional route should start with the agent’s business scope and operating boundary. The agent needs grounded data, approved actions, user-aware permissions, and a clear escalation path before it can safely assist with a Fusion workflow.

Question 2

Topic: Tool invocation

An agent can create a supplier-service request, but the workflow requires supplier ID, issue category, priority, and requester authorization. Which implementation pattern is strongest?

• A. Let the agent call the action as soon as the user mentions a supplier
• B. Store missing fields in the prompt and let the back-end infer them
• C. Disable workflow validation because the agent already understands context
• D. Build a guarded tool call that validates required fields, checks authorization, and returns safe errors

Best answer: D

Explanation: Tool calls that change business data need input validation, authorization checks, and predictable error handling. The agent should collect missing fields and confirm risky changes instead of sending incomplete or unauthorized requests.

Question 3

Topic: Grounding

Users receive different answers because the agent retrieves both a current implementation guide and a retired process document. What should the developer fix?

• A. The source selection, ranking, and exclusion rules so retired guidance is not treated as current
• B. The background color of the chat panel
• C. The user display name format only
• D. The number of examples in the welcome message

Best answer: A

Explanation: Grounding quality depends on the sources the agent can retrieve and how it ranks them. Retired or superseded content should be excluded, downgraded, or clearly labeled so the agent does not present stale process guidance as current.

Question 4

Topic: Context resolution

A user asks, “Can you update it?” after discussing two open service requests. What should the agent do before taking action?

• A. Update the newest request automatically
• B. Ask the user to confirm the target record, requested change, and authority to make the change
• C. Apply the change to both requests
• D. Ignore the request because pronouns are not supported

Best answer: B

Explanation: Vague references are risky before a business action. The agent should resolve the target, confirm the requested update, and verify that the user is allowed to make the change before invoking a tool.

Question 5

Topic: Security

An agent may summarize sensitive worker data from Fusion HCM. Which control is most important?

• A. Use a single service account with broad access so all summaries work
• B. Remove audit logs to keep responses fast
• C. Enforce least privilege, user-aware data access, secure prompts, and auditability
• D. Put sample worker records directly into public prompt examples

Best answer: C

Explanation: AI agent development does not bypass application security. Sensitive Fusion data should remain protected by identity, role, data-access, prompt, and audit controls.

Question 6

Topic: Integration failure

An external shipment-status API times out while the agent is answering a customer-service request. What is the best response behavior?

• A. Invent a likely shipment status to avoid disappointing the user
• B. Return a clear failure message, avoid exposing secrets, and offer a safe next step or escalation
• C. Display the full stack trace and token values
• D. Mark the shipment as delivered

Best answer: B

Explanation: Production agents need safe failure modes. A timeout should not produce fabricated business facts or leak technical secrets; the user should receive clear next steps.

Question 7

Topic: Testing

Which test set is most useful before releasing an agent that can call Fusion workflow actions?

• A. Only a greeting prompt
• B. Only prompts written by the developer
• C. Only a screenshot test of the chat window
• D. Valid requests, ambiguous requests, unauthorized requests, malformed inputs, tool failures, and edge-case business data

Best answer: D

Explanation: Agent testing should cover normal and adverse paths. Business-action agents must be tested for ambiguity, access failure, bad inputs, downstream errors, and realistic process variations.

Question 8

Topic: Observability

A support team needs to explain why an agent selected the wrong workflow. Which implementation detail helps most?

• A. Traceable logs for prompt context, retrieved sources, chosen tool, input payload, authorization result, and workflow response
• B. A longer welcome message
• C. More icons beside each response
• D. A weekly manual summary written by the support team

Best answer: A

Explanation: Observability makes agent behavior reviewable. Logs should support troubleshooting and governance while still protecting confidential data and secrets.

Question 9

Topic: Lifecycle management

An implementation team updates prompts, tool definitions, and grounding sources each quarter. What should the team maintain?

• A. A single production prompt edited directly by anyone
• B. A policy to never retest the agent after changes
• C. Version control, change review, regression tests, deployment discipline, and rollback planning
• D. A spreadsheet of informal comments only

Best answer: C

Explanation: AI agent artifacts should be managed like application assets. Versioning, review, testing, controlled deployment, and rollback planning reduce regression and audit risk.

Question 10

Topic: Human oversight

An agent can recommend a high-impact supplier-bank-detail update. What safeguard should the developer design?

• A. Allow the update whenever the user sounds confident
• B. Save the new bank detail only in the conversation transcript
• C. Hide the action from audit reports
• D. Require strong authorization, validation, approval, and audit trail before the change is submitted

Best answer: D

Explanation: High-risk business changes need stronger controls than a natural-language request. Authorization, validation, approval, and auditability help prevent fraud and data-entry errors.

Question 11

Topic: Response design

An agent gives a correct answer but does not show which source or business rule supported it. What improvement is most useful?

• A. Add grounded citations or source context where appropriate and explain the decision rule clearly
• B. Make every answer shorter than five words
• C. Remove all business terminology
• D. Replace the answer with a generic disclaimer

Best answer: A

Explanation: Users need to understand why an agent answer is reliable. Grounded source context and decision-rule language make the response easier to verify and act on.

Question 12

Topic: Route fit

Which task best matches preparation for Oracle 1Z0-1155-2?

• A. Replacing a laptop battery
• B. Building and governing Fusion AI Agent Studio agents with grounded data, safe tool calls, integration handling, testing, and lifecycle controls
• C. Memorizing only Java syntax
• D. Configuring a standalone network switch

Best answer: B

Explanation: The Rel 26-2 developer professional route is centered on Fusion AI Agent Studio implementation. Strong preparation emphasizes agent design, grounding, tools, security, integrations, testing, and operations.

Quick Cheat Sheet

Related Oracle Routes

• Oracle Fusion AI Agent Studio Foundations Associate
• Older 1Z0-1155-1 replacement note
• Oracle Fusion Applications AI hub
• Oracle certification practice hub

Official Oracle Reference

• Oracle Certification
• Oracle 1Z0-1155-2 exam page