PracticeVideosBlogFlashcardsEspañol
All Practice Exams

100+ Free Oracle Agentic AI Foundations Practice Questions

Prepare for the Oracle Agentic AI Foundations Associate (1Z0-1157-26) exam with instant access — no signup required.

✓ No registration✓ No credit card✓ No hidden fees✓ Start practicing immediately
100+ Questions
100% Free

Loading practice questions...

Same family resources

Explore More Oracle Certifications

Continue into nearby exams from the same family. Each card keeps practice questions, study guides, flashcards, videos, and articles in one place.

Oracle SQL

Practice Questions1 video1 blog

Oracle DBA

Practice Questions

OCI AI Foundations

Practice Questions

OCI Architect

Practice Questions

OCI Foundations

Practice Questions

Oracle Java SE 21 Developer

Practice Questions

More From This Family

Videos and articles for deeper review.

VideoOracle SQL 1Z0-071 Exam Guide 2026: Active Status, Format, and Study PlanArticleOracle SQL 1Z0-071 Exam Guide 2026: Active Status, Format, and Study Plan15 min read
2026 Statistics

Key Facts: Oracle Agentic AI Foundations Exam

40

Exam Questions

Oracle University

60 min

Exam Duration

Oracle University

65%

Passing Score

Oracle University

Free

Exam Fee

Oracle University

6

Learning Path Modules

Oracle MyLearn

1Z0-1157-26

Exam Code

Oracle University

As of mid-2026, Oracle University lists Agentic AI Foundations Associate exam 1Z0-1157-26 as a free, 40-question, 60-minute multiple-choice exam with a 65% passing score. Prep should cover agent fundamentals (LLM+tools+loop, CoT, ReAct, guardrails), LangChain agent patterns, Model Context Protocol, the OpenAI Responses API and Agents SDK, OCI enterprise agent platform capabilities, and Oracle AI Database agent features such as Vector Search, Private Agent Factory, Select AI Agent, and Autonomous AI Database MCP Server.

Sample Oracle Agentic AI Foundations Practice Questions

Try these sample questions to test your Oracle Agentic AI Foundations exam readiness. Each question includes a detailed explanation. Start the interactive quiz above for the full 100+ question experience with AI tutoring.

1Which combination best describes the core architecture of a modern AI agent?
A.A large language model, external tools, and a control loop that pursues a goal
B.A static decision tree, a relational database, and a nightly batch job
C.A single prompt template with no ability to call external functions
D.A rule engine that only returns predefined FAQ answers
Explanation: Foundational agent designs combine an LLM (reasoning and language), tools (actions and data access), and a loop that observes results and continues until the goal is met or limits are hit. Static trees, single-turn prompts, and pure FAQ engines lack that goal-directed tool loop.
2What primarily distinguishes a goal-directed AI agent from a simple chatbot?
A.The agent plans and takes multi-step actions toward an objective, not only generating conversational replies
B.The agent can only answer questions about its own training data cutoff
C.The chatbot always uses tools while the agent never does
D.The agent must run offline without any network access
Explanation: Goal-directed agents decompose objectives and use iterative reasoning and tools to make progress. Chatbots mainly produce responses; they may not plan multi-step tool use to complete a task. The other options reverse tool usage, invent offline requirements, or confuse agents with training-data limits.
3What does Chain-of-Thought (CoT) prompting encourage a model to do?
A.Produce intermediate reasoning steps before the final answer
B.Call every available tool in parallel without reading results
C.Skip reasoning and return only a one-word classification
D.Fine-tune model weights on enterprise documents overnight
Explanation: Chain-of-Thought prompts the model to write intermediate reasoning, which often improves multi-step problem solving. It is a reasoning pattern, not tool orchestration, forced one-word answers, or fine-tuning.
4In the ReAct pattern, what does the agent interleave?
A.Reasoning traces and tool actions based on observations
B.Only SQL statements and CSV exports
C.Model pretraining and full weight fine-tuning
D.User interface themes and CSS stylesheets
Explanation: ReAct (Reason + Act) interleaves natural-language reasoning with tool actions and incorporates observations from those tools into the next thought. It is not limited to SQL/CSV, training, or UI styling.
5Why do production agent systems use layered guardrails?
A.To apply multiple complementary safety and policy checks rather than relying on a single filter
B.To force the agent to ignore all user instructions permanently
C.To replace the need for any logging or monitoring
D.To guarantee zero latency on every tool call
Explanation: Layered guardrails combine checks such as input filtering, tool allowlists, output moderation, and policy validators so failures in one layer can be caught by another. They do not eliminate monitoring needs, ban all instructions, or guarantee latency.
6An agent is asked to book a meeting only if the calendar is free. Which design best matches goal-directed agent behavior?
A.Check availability with a calendar tool, then book only when free, stopping or reporting if not
B.Always invent a booking confirmation without checking any system
C.Return a generic essay about calendar history without taking action
D.Disable tools and answer from parametric memory only
Explanation: Goal-directed agents use tools to verify preconditions and only perform side-effecting actions when conditions hold. Hallucinated confirmations, essays, or tool-less parametric answers fail the operational goal.
7How does an observation from a tool call typically influence the next agent step?
A.It is fed back into the loop so the model can reason about real results and choose the next action
B.It permanently freezes the model weights for that session
C.It is discarded so the model never sees tool output
D.It always ends the session with a failure, even if successful
Explanation: In the LLM+tools+loop design, tool results become observations that condition the next thought and action. They do not freeze weights, get discarded by design, or force failure when successful.
8Which statement correctly contrasts Chain-of-Thought with ReAct?
A.CoT emphasizes multi-step reasoning text; ReAct interleaves reasoning with tool actions and observations
B.CoT always requires tools; ReAct forbids tools
C.CoT is only for images; ReAct is only for audio
D.They are identical names for full model fine-tuning
Explanation: CoT focuses on intermediate reasoning steps. ReAct adds the act/observe cycle with tools. They are complementary reasoning patterns, not modality exclusives or fine-tuning methods.
9Which is an example of a layered guardrail strategy for an enterprise agent?
A.Input policy filter, tool allowlist, sandboxed execution, and output moderation applied together
B.Removing authentication so every user can run every tool
C.Logging nothing so attacks leave no trail
D.Using a single free-text system prompt with no other controls
Explanation: Defense in depth stacks multiple controls across input, tools, execution environment, and outputs. Removing auth, disabling logs, or relying on one prompt alone is the opposite of layered guardrails.
10Why might an agent loop include an explicit stop condition?
A.To prevent unbounded tool calls and control cost, latency, and risk
B.To ensure the agent never returns an answer to the user
C.To force infinite recursion for better accuracy
D.To disable guardrails after the first step
Explanation: Stop conditions (max steps, goal achieved, budget limits) keep agent loops safe and economical. Infinite recursion and disabling guardrails increase risk; never answering fails the product goal.

About the Oracle Agentic AI Foundations Exam

The Oracle Agentic AI Foundations Associate exam (1Z0-1157-26) validates foundational skills for designing and building AI agents. Oracle's learning path covers six modules: Introduction to AI Agents, LangChain for AI Agents, Introduction to MCP, OpenAI Responses API and Agents SDK, Agentic AI for Enterprises on OCI, and Agentic AI for Oracle AI Database. The certification exam is free and targets learners who already have basic LLM familiarity, functional Python, and recommended basic OCI knowledge.

Assessment

40 multiple-choice questions

Time Limit

60 minutes

Passing Score

65%

Exam Fee

Free ($0) (Oracle University)

Oracle Agentic AI Foundations Exam Content Outline

~17%

Introduction to AI Agents

Understand how AI agents combine an LLM, tools, and a control loop to pursue goals; recognize Chain-of-Thought and ReAct patterns; and apply layered guardrails for safer agent behavior.

~17%

LangChain for AI Agents

Build and run agents with LangChain Expression Language (LCEL), tool schemas, agent.invoke, and the tool-call parse/execute loop used in production agent apps.

~16%

Introduction to MCP

Explain Model Context Protocol architecture, how MCP servers expose tools, resources, and prompts, why MCP decouples clients from tool implementations, and how OCI Usage MCP fits enterprise examples.

~17%

OpenAI Responses API and Agents SDK

Contrast the Responses API with the Agents SDK, use function calling, design multi-agent handoffs, apply guardrails, and interpret tracing for observability.

~16%

Agentic AI for Enterprises / OCI Enterprise AI

Recognize enterprise agent platform needs: hosted endpoints, scaling, memory and sessions, sandboxed tool execution, logging, and integrations on OCI.

~17%

Agentic AI for Oracle AI Database

Connect agent workflows to Oracle AI Database capabilities including Vector Search, Private Agent Factory, Select AI Agent, and the Autonomous AI Database MCP Server.

How to Pass the Oracle Agentic AI Foundations Exam

What You Need to Know

  • Passing score: 65%
  • Assessment: 40 multiple-choice questions
  • Time limit: 60 minutes
  • Exam fee: Free ($0)

Keys to Passing

  • Complete 500+ practice questions
  • Score 80%+ consistently before scheduling
  • Focus on highest-weighted sections
  • Use our AI tutor for tough concepts

Oracle Agentic AI Foundations Study Tips from Top Performers

1Memorize the agent triad: LLM + tools + control loop working toward an explicit goal—not just single-turn chat completion.
2Be able to contrast Chain-of-Thought (reasoning steps) with ReAct (interleaved reasoning and tool actions).
3Practice reading tool schemas and predicting the parse → execute → observe → continue loop in LangChain-style agents.
4Know MCP's three exposure types—tools, resources, and prompts—and why a protocol layer decouples hosts from tool servers.
5Distinguish when to use a lower-level Responses/function-calling path versus a higher-level Agents SDK with handoffs and guardrails.
6Map Oracle AI Database agent pieces: Vector Search for retrieval, Select AI Agent for NL-to-SQL/agent patterns, Private Agent Factory for enterprise agent construction, and ADB MCP Server for protocol-based tool access.
7Use timed 40-question practice at 60 minutes so free-exam pacing feels natural on test day.

Frequently Asked Questions

What are the official exam facts for Oracle Agentic AI Foundations Associate?

Oracle University lists exam 1Z0-1157-26 as a free multiple-choice certification with 40 questions, a 60-minute time limit, and a 65% passing score. Register through Oracle's exam product page for Agentic AI Foundations Associate.

Is the 1Z0-1157-26 certification exam really free?

Yes. Oracle University promotes the Agentic AI Foundations Associate certification exam as free. Always confirm the current price on Oracle's official exam page before scheduling, but Oracle's published offering for this track is $0 exam fee.

What prerequisites does Oracle recommend?

Oracle's course framing recommends basic familiarity with large language models, functional Python skills, and basic OCI knowledge. These are recommended foundations rather than formal gatekeeping requirements for registering for the free exam.

What six modules does the learning path cover?

The Become an OCI Agentic AI Foundations Associate path covers Introduction to AI Agents, LangChain for AI Agents, Introduction to MCP, OpenAI Responses API and Agents SDK, Agentic AI for Enterprises (OCI Enterprise AI), and Agentic AI for Oracle AI Database.

How does this differ from OCI AI Foundations?

OCI AI Foundations focuses on broader AI/ML/DL concepts and OCI AI services. Agentic AI Foundations Associate focuses specifically on building and operating AI agents—frameworks like LangChain, protocols like MCP, agent SDKs, enterprise hosting, and Oracle AI Database agent features.

What is the best study approach for 1Z0-1157-26?

Start with agent fundamentals and ReAct/CoT patterns, then practice LangChain tool-calling loops and MCP architecture. Finish with OpenAI Agents SDK handoffs/guardrails and Oracle-specific topics (OCI hosted agents, Vector Search, Select AI Agent, Private Agent Factory, ADB MCP). Use timed 40-question sets to match the free exam pacing.