1Z0-1145-1 Exam Guide: What Should You Actually Study?

If you are preparing for 1Z0-1145-1 in September 2026, the difficult part is not finding information. There is plenty of it. The harder job is deciding which parts of Oracle Fusion AI Agent Studio deserve serious study and which details can wait. The exam is small enough to look approachable, but the product model behind it has several connected pieces that are easy to mix up.

What Is the 1Z0-1145-1 Exam?

The Oracle Fusion AI Agent Studio Foundations Associate – Rel 1 exam, code 1Z0-1145-1, is currently listed by Oracle as a 60-minute, 40-question exam with a 65% passing score. Oracle lists it as a no-cost exam.

The certification is centered on Oracle Fusion AI Agent Studio, so preparation should stay close to the way the product actually works. Spending most of your time on general AI theory, broad large-language-model concepts, or unrelated Oracle Cloud services is unlikely to be the best use of the limited study window.

Oracle’s own training material describes the exam-topics module as a way to review the key topics after completing the Fusion AI Agent Studio course.

That gives us a sensible starting point: use the official objectives as the boundary, then study the product relationships behind those objectives.

What Does 1Z0-1145-1 Actually Test?

The exam topics cover areas including core AI Agent Studio capabilities, configuring and deploying agents, customizing out-of-the-box agents, evaluating and testing agents, implementing security, and monitoring agents. Third-party preparation pages currently reproduce these six areas, but Oracle’s own exam information should remain the authority when deciding what belongs in your preparation.

The six areas are easier to study when viewed as parts of one lifecycle rather than six unrelated chapters.

An agent needs a role and capabilities. Topics shape what it can handle. Tools give it access to information or actions. Instructions guide how it behaves. An agent or group of agents is assembled into an agent team, which can then be tested, published, monitored, and evaluated.

Oracle’s current 26C documentation reflects this product model. An agent team is the deployable component, while agents perform specialized roles within the team. Topics define areas of expertise and boundaries, tools provide additional capabilities, and instructions establish natural-language rules and guardrails.

That relationship is much more useful for exam preparation than memorizing six isolated definitions.

Agent vs. Agent Team

This is one distinction worth getting right early.

An agent uses an LLM to reason, create action plans, interact with users, and perform tasks. Oracle identifies different agent types, including user-proxy, supervisor, and specialist or utility agents. An agent must be part of an agent team to be deployed.

The agent team provides the larger structure around the work. It can contain one agent or several agents and defines the overall flow for a business task.

If a question describes several specialized agents working together under a coordinating structure, do not reduce the scenario to “which agent does this?” The relationship between the individual agent and the deployable team is part of the product model you need to recognize.

Oracle also documents workflow-style agent teams, where tasks are represented by connected nodes and each node performs a defined function. These workflows are intended for deterministic, rule-based orchestration.

Topics, Tools, and Instructions Need to Be Studied Together

This is another area where reading definitions one after another can create unnecessary confusion.

Consider a benefits agent that answers employee questions.

The topic establishes the area of expertise and helps define the boundaries of the conversation. A topic might cover medical benefits or retirement plans.

The tool gives the agent a capability it needs to complete a task, such as retrieving information from a business object or searching documents.

The instructions tell the agent how to behave within that context. Oracle describes instructions as natural-language rules or conditions that become part of the prompts sent to the underlying LLM.

A question can therefore describe a business requirement without naming the component you need.

Suppose an agent must retrieve a customer’s current Fusion record. The question may be testing whether you recognize when a Business Object tool is appropriate. If it needs to communicate with an external service, an External REST tool may be the relevant capability. Oracle’s current documentation describes Business Object tools as a way for agents to retrieve, update, create, or delete Fusion business-object records, while External REST tools can connect to internal or third-party services and public APIs.

I would spend more time comparing these situations than memorizing a long tool list.

Where Should You Spend Most of Your Study Time?

There is no need to invent percentage weights for the six exam domains when Oracle does not publish them. A better approach is to prioritize the areas where several exam concepts intersect.

Start With the Core Product Model

Before working through practice questions, make sure you can explain the relationship between:

• Agent teams
• Agents
• Topics
• Tools
• Instructions
• Testing and evaluation
• Deployment
• Monitoring

You do not need to reproduce Oracle documentation word for word. You should be able to read a short business scenario and identify which part of AI Agent Studio is being described.

Oracle’s current documentation also shows that the platform now includes workflows, human approval steps, external REST connections, connectors, document tools, MCP tools, and other capabilities.

That does not mean every new feature deserves equal study time. Focus on functionality that connects directly to the current exam objectives.

Give Testing and Evaluation More Attention Than Their Names Suggest

Testing and evaluation are particularly easy to blur together.

AI Agent Studio supports iterative testing in the Playground, where agents can be modified and tested without deploying the changes to production. Oracle’s current documentation also describes evaluation sets containing test questions, expected responses, and metrics.

An evaluation set can compare expected and actual responses and apply metrics such as correctness. Oracle also documents measures including latency, token usage, groundedness, answer relevance, and context relevance.

A useful preparation exercise is to separate the purposes:

Testing asks whether the configuration behaves correctly under a given scenario.

Evaluation gives you a structured way to measure agent performance against expected outcomes.

Monitoring concerns the behavior and performance of agents over actual runtime activity.

Oracle’s monitoring documentation shows session status, token usage, tool calls, step durations, and detailed traces.

If a question describes an agent that is already running and asks about ongoing performance or runtime behavior, monitoring is a different context from preparing an evaluation set before deployment.

You can surely appreciate the advantages of this distinction.

Security Should Not Be Treated as a Separate Afterthought

Security appears as its own exam area, but it also connects to how tools and agents access Fusion data.

Oracle’s current AI Agent Studio documentation states that Business Object tools can be configured around the business objects and fields an agent can access, while agents follow native Fusion security and role-based access controls.

Oracle also documents specific access requirements for AI Agent Studio, including predefined duties and optional permissions for external REST tools, channels, scheduling, and monitored agent records.

For preparation, pay attention to scenarios involving who can access information, what an agent is allowed to do, and how permissions affect the agent’s operation.

That is more useful than treating “security” as a collection of generic cybersecurity terms.

Do You Need Hands-On Experience With AI Agent Studio?

You do not need to turn this certification into a full implementation project.

A small amount of practical exposure can make the product model much easier to follow, particularly if Fusion Applications is new to you. Even working through a simple example involving an agent, topic, tool, instruction, test query, and deployment flow can remove a lot of ambiguity.

Oracle’s current getting-started documentation follows a similar progression: define tools, create topics and agents, build the agent team or workflow, test it, and deploy it.

If you already work with Fusion Applications, you may find the business context familiar even if AI Agent Studio itself is new. If Fusion is unfamiliar, give yourself more time with Oracle’s terminology and interface structure before moving heavily into question practice.

How Difficult Is 1Z0-1145-1?

The 60-minute, 40-question format does not tell you much about the intellectual difficulty of the exam.

A candidate who already understands Fusion Applications and has seen AI Agent Studio will approach the terminology differently from someone encountering Business Objects, agent teams, topics, tools, and Oracle-specific security concepts for the first time.

The potential difficulty comes from choosing between related capabilities in a business scenario.

A question may not simply ask what a Business Object tool is. It can describe a requirement and expect you to identify the appropriate mechanism. The same applies to agent types, tools, testing, evaluation, and monitoring.

That is why broad familiarity with the product can matter more than the raw number of pages in your study notes.

How Should You Prepare for the Exam?

Use Oracle’s official material as the foundation. Then turn the exam objectives into a personal checklist.

For each topic, ask yourself whether you can do three things:

Recognize it.
Can you identify the component or capability when it appears in a scenario?

Differentiate it.
Can you tell it apart from a closely related component?

Apply it.
Can you choose the appropriate capability when the question describes a business requirement?

That third level is where practice questions become useful.

For example, you might know that AI Agent Studio has Business Object and External REST tools. That is one level of knowledge. Being given a scenario involving Fusion business records and an external logistics platform and choosing the appropriate tool requires a different type of preparation.

Current practice resources also show that scenario-style questions are being built around areas such as agent customization, evaluation data, tool selection, security, and performance metrics. These resources constitute third-party content; they should be regarded as supplementary study aids and utilized once a foundational understanding has been established.

Use them to expose gaps, not as a substitute for Oracle’s documentation.

Practice Questions Have a Specific Job

Practice questions are most useful after you have established the product foundation.

They can reveal a gap that reading alone may not expose:

You recognize the term, but can you choose the right capability when the wording changes?

That is particularly relevant to 1Z0-1145-1 because several product components have related roles.

A sensible preparation cycle is:

Official objectives and documentation → study the product relationships → answer practice questions → review mistakes → return to the relevant Oracle documentation.

The review stage matters. If you miss a question about evaluation metrics, simply recording the correct option does not solve the problem. Go back and establish why that metric belongs to evaluation, monitoring, or another part of the product.

If you want an additional question-based resource while reviewing the official objectives, you can see the current 1Z0-1145-1 preparation options at Leads4Pass.

The resource should supplement your preparation rather than replace the Oracle material.

When Are You Ready to Take 1Z0-1145-1?

Readiness is less about reaching an arbitrary number of practice questions and more about what happens when the wording changes.

Before booking the exam, check whether you can comfortably explain:

• What an agent team does compared with an individual agent
• How topics, tools, and instructions relate to one another
• When a Business Object tool fits a requirement
• When an External REST tool is more appropriate
• How out-of-the-box agents can be customized
• Why testing, evaluation, and monitoring are separate activities
• What evaluation sets are used for
• What types of metrics can be assessed
• How deployment fits into the agent lifecycle
• How security and access controls affect agent behavior

You should also be able to review a missed practice question without relying on the answer choice itself to trigger your memory.

If you consistently need to eliminate options by recognizing familiar wording rather than understanding the scenario, there is probably still a product-level gap to fix.

Conclusion

The most efficient way to prepare for 1Z0-1145-1 in September 2026 is to study the exam as a product workflow rather than as a vocabulary list.

Start with Oracle’s current objectives. Learn how AI Agent Studio’s agents, agent teams, topics, tools, and instructions fit together. Spend deliberate time on testing, evaluation, monitoring, deployment, customization, and security. Then use practice questions to find where your understanding breaks down under different scenarios.

You do not need to know every corner of AI Agent Studio. You need enough command of the current product model to recognize what a business requirement is asking for and why one capability fits better than another.

That is the study time most likely to pay off when you sit down with the 1Z0-1145-1 exam in front of you.