UiPath-AAAv1 sample questions

Question 1

You are configuring a Service Task in UiPath Maestro to invoke an Agent. The business process requires the agent to analyze a document and return a 'Risk Score' which determines the next path in the BPMN gateway. Which configuration is mandatory for this gateway logic to work?

Accepted Answer

The Service Task must output the 'Risk Score' to a process variable, and the Gateway must define conditions based on that variable (e.g., RiskScore > 80).

Answer

The Service Task must be set to 'Fire and Forget'.

Answer

The Gateway must be configured as a Parallel Gateway.

Answer

The Agent must be configured to send an email to the process owner.

Question 2

Which of the following is a key differentiator between a 'Deterministic' evaluation and a 'Model-graded' evaluation for an AI Agent?

Accepted Answer

Deterministic evaluations check for strict conditions (e.g., 'Is output valid JSON?'), while Model-graded evaluations assess subjective qualities (e.g., 'Is the answer helpful?').

Answer

Deterministic evaluations are slower than Model-graded evaluations.

Answer

Deterministic evaluations use an LLM to judge quality; Model-graded evaluations use RegEx.

Answer

Model-graded evaluations require human intervention for every case.

Question 3

You are creating a prompt for an agent to classify customer emails into categories: 'Complaint', 'Inquiry', or 'Feedback'. You want to ensure the agent outputs ONLY the category name and nothing else. Which Prompt Engineering technique should you apply?

Accepted Answer

Add a constraint instruction: 'Output ONLY the category name. Do not include explanations or other text.'

Answer

Increase the temperature setting of the model.

Answer

Ask the agent to 'Think about it carefully'.

Answer

Use a Zero-shot prompt with just the email content.

Question 4

In UiPath Studio Web, you are creating a new 'Tool' for an agent to check inventory. You have defined the Input Schema to accept a 'ProductCode'. Why is it critical to also provide a clear, descriptive 'Description' for the tool itself?

Accepted Answer

The Description is used by the Agent's reasoning engine (LLM) to decide WHEN and WHY to use this specific tool.

Answer

The Description determines the API endpoint URL.

Answer

The Description is displayed to the end-user in the chat window.

Answer

It is not critical; the agent analyzes the workflow code to understand the tool.

Question 5

Select TWO primary benefits of 'Agentic Automation' compared to traditional 'Rule-Based Automation'. (Select TWO)

Accepted Answer

Dynamic decision-making and planning capabilities (reasoning) to achieve a goal.

Accepted Answer

Ability to handle unstructured data and ambiguous instructions without strict pre-programming.

Answer

Faster execution speed for high-volume, strictly structured transactional data.

Answer

Lower computational cost and zero reliance on external models.

Question 6

A financial services firm is implementing an automated loan approval process. The process requires analyzing unstructured borrower emails, extracting intent, and conditionally retrieving credit reports from a legacy mainframe system via an existing RPA automation. However, the final approval decision involves weighing qualitative factors from the email against the quantitative credit score, a task that varies significantly per case. Which architectural approach best leverages the synergy between Agentic AI and traditional RPA for this scenario?

Accepted Answer

Implement an Agentic Workflow where an AI Agent acts as the orchestrator: it analyzes the email, invokes the RPA robot as a 'tool' to fetch the credit score, and then synthesizes both data points to recommend a decision.

Answer

Use a traditional RPA robot to process the emails using RegEx and trigger the mainframe lookup, then forward all cases to a human for the final decision.

Answer

Deploy a single Generative AI model to handle both the email analysis and the mainframe interaction directly via API calls.

Answer

Configure UiPath Maestro to run a BPMN process that strictly alternates between an RPA robot for email reading and an AI Agent for mainframe data entry.

Question 7

You are assessing a complex Customer Support process for potential Agentic Automation. The process involves technicians diagnosing hardware issues based on vague customer descriptions, consulting dynamic PDF technical manuals, and scheduling on-site visits. The legacy scheduling system has no API but has a reliable web interface. Why is this process a prime candidate for an Agentic solution rather than pure Generative AI or traditional RPA?

Accepted Answer

Because the process requires probabilistic reasoning to diagnose issues (Agent strength) combined with the ability to execute actions in a system via tools (Agent capability), which pure GenAI lacks.

Answer

Because traditional RPA is incapable of scraping text from web interfaces, requiring an Agent's vision capabilities.

Answer

Because Generative AI cannot read PDF files and RPA cannot access web interfaces without APIs.

Answer

Because Agentic Automation is cheaper to license than the combination of Document Understanding and Orchestrator Robots.

Question 8

When designing an Agent Blueprint for a 'Travel Concierge' agent, you need to define the System Prompt. Which of the following components should be included in the System Prompt to ensure the agent behaves consistently and securely?

Accepted Answer

Operational constraints (e.g., 'Do not book flights costing over $2000 without escalation').

Accepted Answer

The agent's persona (e.g., 'You are an experienced travel agent').

Answer

The API key for the flight booking system.

Answer

The specific travel dates requested by the current user.

Question 9

A developer is configuring Context Grounding for an HR Policy Agent. The agent must answer employee questions based on a specific set of 50 PDF documents stored in a UiPath Storage Bucket. The developer wants to ensure the agent cites the specific policy section when answering. Which configuration step is essential to achieve this RAG (Retrieval-Augmented Generation) pattern?

Accepted Answer

Create an Index in the Context Grounding service, ingest the files from the Storage Bucket, and map this Index to the Agent's context property.

Answer

Convert all PDFs to text and paste the content into the Agent's System Prompt.

Answer

Use the 'Read PDF Text' activity in a loop and pass the output variable to the 'Generate Chat Response' activity.

Answer

Train a custom LLM model using the 50 PDFs and deploy it to AI Center.

Question 10

In a 'Claims Processing' agent, you need to implement a Human-in-the-Loop (HITL) pattern. The agent should automatically approve claims under $500 but must escalate claims over $500 or claims with 'Suspicious' fraud flags to a human manager. What is the correct logic implementation for this escalation?

Accepted Answer

Configure a 'Create Action' task within the agent's toolset and instruct the agent to invoke this tool when the claim amount > $500 OR fraud_flag equals 'Suspicious'.

Answer

Use a deterministic 'If' activity in the underlying RPA workflow to check the amount before the Agent is even called.

Answer

The agent automatically detects high-value transactions based on historical data and pauses execution without configuration.

Answer

Instruct the Agent in the System Prompt to 'Stop and output error' if the amount is > $500.

UiPath-AAAv1 UiPath Certified Professional Agentic Automation Associate Practice Questions

Create an account to continue.

Create an account to continue.

Related exams