UUS Practice Tests⌕ Search
IT Certifications · 1Z0-1145-1

Oracle Fusion AI Agent Studio Foundations Associate - Rel 1 Exam Practice Test

The Oracle Fusion AI Agent Studio Foundations Associate - Rel 1 Exam covers foundational decisions involved in building and operating business agents. Practice topics include selecting tools for tasks, handling capability outcomes, protecting credentials and personal data, and checking that agents use appropriate business information.

Exam overview

Exam providerOracle
Exam code1Z0-1145-1
CategoryIT Certifications

About the Oracle Fusion AI Agent Studio Foundations Associate - Rel 1 Exam

Agent designers need to distinguish tasks suited to a language model from those requiring a tool or action, such as retrieving current information or invoking an external service. They also need to plan for successful and failed capability calls so workflows can respond appropriately. Operational work includes comparing configurations against the same tasks, monitoring behavior after changes, and recording useful diagnostic details while respecting privacy. When responses rely on retrieved documents, traceability helps connect generated answers to source material. Access checks help confirm that an agent can reach only the intended business data. These skills support design and evaluation decisions across an agent’s lifecycle, from configuring its interactions to investigating unexpected results in use.

Sample Practice Questions

Review a selection from this exam's published US Practice Tests practice bank.

Question 1

An agent workflow needs to distinguish a successful capability call from a failed one. Which result design is most helpful?

  1. AReturn an explicit success or error status along with relevant result details
  2. BReturn an empty response for both success and failure
  3. CPut all outcomes into an unstructured paragraph
  4. DTreat every completed call as successful regardless of its result
Show answerCorrect answer: A

A clear status lets later workflow logic handle success and failure differently, while result details can support recovery or an informative response.

Question 2

A team needs to determine whether a deployed agent is behaving differently after a configuration update. Which practice is most useful?

  1. ACompare outcomes from representative requests before and after the update
  2. BChange several unrelated settings at the same time and rely on memory
  3. CReview only the agent's display name
  4. DWait for users to report every possible issue
Show answerCorrect answer: A

A controlled comparison using representative requests helps identify behavioral changes attributable to the update. Changing many variables at once makes diagnosis harder.

Question 3

A capability requires a credential to call an external service. Which approach best protects that credential?

  1. AUse a managed connection or secure credential store rather than embedding the secret in instructions
  2. BPlace the credential in a public description so maintainers can find it
  3. CInclude the credential in every user prompt
  4. DStore the credential in a workflow label
Show answerCorrect answer: A

Managed connections or secure credential stores are designed to control and protect secrets. Embedding secrets in prompts or visible configuration text risks disclosure.

Question 4

Which information is most useful to capture in an operational log for a business agent, subject to applicable privacy rules?

  1. AThe interaction or action outcome, relevant timestamps, and enough context to investigate failures
  2. BEvery available personal data field, whether relevant or not
  3. COnly successful actions, omitting errors and rejected requests
  4. DThe agent's hidden instructions in a public user-facing response
Show answerCorrect answer: A

Useful operational logs support troubleshooting and oversight by recording relevant events and outcomes. Logging should be proportionate and follow privacy and retention requirements rather than collecting unnecessary data.

Question 5

Why would an agent use a tool or action rather than answer entirely from its language model?

  1. ATo perform a defined operation, such as retrieving current business data or submitting a request
  2. BTo make the agent's responses independent of all permissions
  3. CTo eliminate the need to validate inputs
  4. DTo ensure the agent can access every application automatically
Show answerCorrect answer: A

Tools let an agent interact with systems or perform bounded operations that a text response alone cannot carry out. Tool use remains subject to configuration, permissions, and validation.

Question 6

A team is comparing two agent configurations for a common task. Which evaluation setup gives the clearest comparison?

  1. AUse the same representative test requests and success criteria for both configurations
  2. BUse unrelated requests and different success criteria for each configuration
  3. CJudge only by which configuration has more settings
  4. DEvaluate one configuration with users and the other without any tests
Show answerCorrect answer: A

Holding test requests and success criteria constant makes the comparison more meaningful. Different inputs or criteria can confound the results.

Question 7

When an agent uses retrieved documents to draft an answer, what is a useful way to improve the answer's traceability?

  1. AIdentify the supporting documents or passages used, when the interface supports it
  2. BRemove all references so the answer reads more smoothly
  3. CPresent every retrieved document as equally relevant
  4. DClaim that retrieval guarantees the answer is correct
Show answerCorrect answer: A

Showing which retrieved material supports an answer helps users check its basis. Retrieval can aid grounding, but relevance and correctness still need evaluation.

How to use this practice test

Access is free. Each session presents up to 20 random questions from the available 50-question practice bank and shows results afterward. Use the questions to check your reasoning about tool use, capability outcomes, evaluation, privacy, logging, and data access; revisit any missed scenarios and identify the design decision they test.

Preparation tips

  • For each scenario, decide whether the agent should answer from its model or call a tool, especially when the task depends on current external information.
  • Practice mapping successful and failed capability results to distinct workflow responses, including what the agent should do after a failed call.
  • Compare agent configurations using the same tasks and evaluation conditions so behavior differences are easier to attribute.
  • Review logging choices for diagnostic value, and exclude or protect personally identifiable information when it is not needed.
  • Trace retrieved answers to their supporting documents and check that the agent's data access is limited to the intended business information.
Official sourceOfficial exam page ↗