AB-410 Business Application Logic And Automation Practice Test 10 — 15 Questions

Question 1

You are designing a Power Automate cloud flow that must process a large batch of records from a Microsoft Dataverse table. The flow will be triggered manually and needs to iterate through all records, perform an action on each, and handle errors gracefully. You want to ensure the flow is efficient and does not exceed service limits. Which two actions or configurations should you use? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Enable 'Concurrency Control' on the 'Apply to each' loop and set a degree of parallelism greater than 1.. To process a large batch efficiently, retrieve only the needed records with a filtered 'List rows' action, then iterate with 'Apply to each'. Enabling Concurrency Control on the loop allows parallel processing, reducing run time. These two configurations optimize performance and help avoid hitting service limits, while the other options either use inefficient patterns or non-existent features.

Answer

Use a 'Do until' loop with a condition that checks if all records have been processed, and retrieve records one at a time.

Answer

Configure the flow to run with a service account that has elevated privileges to avoid throttling.

Answer

Use the 'Batch' action to group records into a single API call.

Question 2

You are designing a Power Automate flow that uses AI Builder to process a batch of documents. You need to handle potential errors when the AI Builder action fails due to unsupported file formats. Which TWO actions or configurations should you implement to ensure robust error handling? (Choose two.)

Accepted Answer

Use a Scope action to group the AI Builder action and configure its run after to handle failures.. To handle errors from unsupported file formats, you should prevent them from reaching the AI Builder action by adding a Condition to check file extensions. Additionally, using a Scope action around the AI Builder action allows you to catch any failures and respond appropriately, such as notifying an admin. These two approaches together provide both proactive and reactive error handling.

Answer

Set the AI Builder action's retry policy to retry on failure.

Answer

Configure the AI Builder action's run after settings to run only on failure and send a notification.

Answer

Enable 'Secure Inputs' on the AI Builder action to prevent errors.

Question 3

You are building a Power Automate flow that uses an AI Builder model to process invoices. The flow must extract data from the invoices and then validate the extracted total against a threshold. Which TWO actions are required to achieve this? (Choose two.)

Accepted Answer

Use the 'Condition' action to compare the extracted total to the threshold.. To extract data from invoices, the AI Builder invoice processing action is required. To validate the total against a threshold, a Condition action is needed to compare the extracted value. Together, these two actions form the core of the validation flow.

Answer

Use the 'AI Builder - Create text with GPT' action to generate a validation message.

Answer

Use the 'Apply to each' action to iterate over multiple invoices.

Answer

Use the 'Parse JSON' action to parse the AI Builder output.

Question 4

You are designing a Power Automate cloud flow that must process a large batch of records. The flow retrieves 10,000 items from a SharePoint list and then updates each item based on a condition. You need to optimize performance and avoid hitting service limits. Which approach should you use?

Accepted Answer

Use the 'Get items' action with a filter query to reduce the dataset, then use 'Apply to each' with concurrency control enabled.. Filtering the SharePoint data with a filter query minimizes the number of items retrieved, and enabling concurrency in 'Apply to each' allows multiple iterations to run in parallel. This reduces overall execution time and helps avoid throttling. It is the most efficient approach for processing a large batch within a single flow.

Answer

Retrieve all items using 'Get items' without a filter, then use a single 'Update item' action outside a loop.

Answer

Use a 'Apply to each' loop with a 'Delay' action inside to throttle requests and avoid rate limits.

Answer

Split the flow into multiple child flows, each processing a subset of items, and trigger them sequentially.

Question 5

You are designing a Power Automate flow that uses an AI Builder custom model to classify support tickets. The model returns a category and a confidence score. You need to route tickets with a confidence score below 0.7 to a human reviewer, and tickets with a score of 0.7 or higher to an automated resolution queue. Which expression should you use in a Condition action to evaluate the confidence score?

Accepted Answer

less(outputs('Classify_ticket')?['body/confidenceScore'], 0.7). The condition must route tickets with confidence below 0.7 to human review. The 'less' function correctly evaluates to true when the score is strictly less than 0.7. Using 'greater' would invert the logic, 'equals' would miss most cases, and 'lessOrEquals' would incorrectly include the boundary value of 0.7, which should go to automation.

Answer

lessOrEquals(outputs('Classify_ticket')?['body/confidenceScore'], 0.7)

Answer

equals(outputs('Classify_ticket')?['body/confidenceScore'], 0.7)

Answer

greater(outputs('Classify_ticket')?['body/confidenceScore'], 0.7)

Question 6

You are designing a Power Automate flow that uses an AI Builder model to extract data from invoices. The flow must handle cases where the model returns a low confidence score for a field. Which approach best ensures that low-confidence extractions are reviewed by a human before being committed to a SharePoint list?

Accepted Answer

Add a condition that checks the confidence score against a threshold; if below, create an approval request in Microsoft Power Automate and wait for the outcome before writing to SharePoint.. To ensure human review of low-confidence extractions, the flow should evaluate the confidence score and, if it is below a threshold, route the document to an approval process. The approval action pauses the flow until a human responds, and only then does the flow proceed to write the data to SharePoint. This pattern guarantees that questionable data is vetted before being committed.

Answer

Configure the AI Builder model to automatically retry extraction up to three times before returning a result.

Answer

Set the AI Builder model to 'strict' mode, which rejects any extraction with a confidence score below 100%.

Answer

Use the 'Create item' action to write all extracted data to SharePoint, then send an email notification to the reviewer with the item link.

Question 7

You are building a Power Automate cloud flow that must call two different Copilot Studio agents in parallel to analyze a customer inquiry, then combine their responses. You need to ensure both agents receive the same input and that the flow waits for both before merging results. (Choose two.)

Accepted Answer

After the Parallel branch, use 'Compose' to combine the outputs from both agent actions.. To call two agents concurrently and merge their results, place both 'Run a Copilot' actions in a Parallel branch so they execute simultaneously, and then use a Compose action after the branch to combine the two outputs. The Parallel branch ensures the flow waits for both to finish, and Compose provides a simple merge point for downstream steps.

Answer

Use a single 'Run a Copilot' action and pass both agents' names in a comma-separated list.

Answer

Configure the second agent to trigger automatically when the first agent finishes by using a 'When a response is received' trigger.

Answer

Set the flow's concurrency control to 1 so the agents do not conflict.

Question 8

You are building a Power Automate flow that uses the AI Builder 'Extract information from invoices' action. The flow processes invoices from multiple vendors, and you need to ensure that the extracted data is stored in a structured format. Which output property of the AI Builder action should you use to access the extracted fields?

Accepted Answer

The 'Response' dynamic content, which contains a JSON object with all extracted fields.. The AI Builder invoice processing action outputs a 'Response' property that contains a JSON object with all extracted fields. This is the structured data you need to store or process further. Other outputs like file content, model ID, or confidence score do not provide the extracted invoice details, so they are not suitable for this requirement.

Answer

The 'Confidence Score' dynamic content, which indicates the accuracy of the extraction.

Answer

The 'File Content' dynamic content, which contains the raw file bytes.

Answer

The 'Model ID' dynamic content, which identifies the AI model used.

Question 9

You are building a Power Automate flow that must call a custom AI Builder model to extract key-value pairs from scanned receipts. The model has been trained and published. You need to pass the receipt image to the model and retrieve the extracted fields. Which action should you add to the flow?

Accepted Answer

Extract information from documents. To use a custom AI Builder document processing model, you must invoke the 'Extract information from documents' action. It is designed to take a document (like a receipt) and return the extracted key-value pairs. Other actions either perform different AI tasks or are for OCR without structure, which would require additional parsing.

Answer

Create text with GPT

Answer

Get text from image

Answer

Predict

Question 10

You are building a Power Apps canvas app that uses AI Builder to classify customer messages into categories. The app must display the predicted category and confidence score. You need to configure the app to call the AI Builder model. Which component should you use?

Accepted Answer

Add the AI Builder model as a data source and use the Predict function in Power Fx.. AI Builder models can be added directly as data sources in canvas apps. The Predict function then invokes the model, returning the predicted category and confidence score. This native integration avoids the overhead of a cloud flow and is the recommended approach for real-time predictions within the app.

Answer

Create a Power Automate cloud flow that calls the model and returns the result, then call the flow from the app.

Answer

Embed the model in a custom connector and call it using the 'Invoke custom connector' action in Power Fx.

Answer

Use the 'AI Builder' connector in Power Apps Studio to send the message to the model via an HTTP request.

Question 11

You need to create a Power Automate flow that processes a large batch of documents using an AI Builder custom model. The flow must process each document independently and in parallel to reduce overall processing time. Which action should you use to achieve parallel processing?

Accepted Answer

Apply to each with concurrency control enabled. To process a batch of documents in parallel, use the 'Apply to each' action with concurrency control enabled. This setting allows multiple iterations of the loop to run simultaneously, up to a specified limit. It is the standard way to achieve parallel processing over a dynamic array in Power Automate, significantly reducing the time to process many documents.

Answer

Parallel branch

Answer

Apply to each

Answer

Do until

Question 12

You are using AI Builder's document processing model to extract invoice total amounts from PDFs. You need to ensure the extracted data is automatically saved to a SharePoint list. What is the most efficient way to achieve this in Power Automate?

Accepted Answer

Use the 'Create item' action and select the dynamic content fields provided by the AI Builder 'Extract information from documents' action.. The AI Builder document processing action returns structured data that is exposed as dynamic content in Power Automate. Using the 'Create item' action and mapping those dynamic content fields directly to SharePoint columns is the most efficient method. It eliminates manual mapping and ensures the flow adapts if the model's schema changes. The other options add unnecessary steps or manual effort.

Answer

Use the 'Create item' action for SharePoint and manually map each field from the document processing output.

Answer

Use the 'Apply to each' action to iterate through the extracted documents and then use 'Create item' with dynamic content from the AI Builder model.

Answer

Export the AI Builder model results to Excel and then use a scheduled flow to import the Excel file into SharePoint.

Question 13

You are designing a Power Automate cloud flow that must process a large batch of records from a Microsoft Dataverse table. The flow will be triggered manually and needs to iterate through all records, perform an action on each, and handle errors gracefully. You want to ensure the flow is efficient and does not exceed service limits. Which two actions or configurations should you use? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Condition control with an expression comparing the confidence score to the threshold.. The Condition control is the correct choice because it evaluates a logical expression and branches the flow accordingly. By comparing the AI Builder confidence score to a threshold, you can route emails to automated responses or human review. Other actions either do not provide branching or are overkill for a simple threshold comparison.

Answer

Switch control with cases for each confidence level.

Answer

Terminate action with a custom status code.

Answer

Apply to each loop with a nested condition.

Question 14

You are building a Power Automate flow that uses an AI Builder model to analyze customer feedback and then triggers a Copilot Studio agent to generate a personalized response. Which TWO actions or configurations are required to enable the flow to call the agent and use its response? (Choose two.)

Accepted Answer

Map the AI Builder output to the agent's input parameters in the 'Invoke an agent' action.. To enable a Power Automate flow to call a Copilot Studio agent and use its response, you must add the 'Invoke an agent' action and map the AI Builder output to the agent's input parameters. These two steps establish the connection and provide the necessary data for the agent to generate a response, which the flow can then use.

Answer

Create a custom connector to the Copilot Studio agent's API.

Answer

Configure the agent to use the 'Generative answers' feature.

Answer

Enable the 'Respond to the user' setting in the agent's authoring canvas.

Question 15

You are building a Power Automate flow that uses an AI Builder document processing model to extract data from invoices. The flow must handle cases where the model's confidence score is below a threshold. Which two actions should you include in the flow? (Choose two.)

Accepted Answer

Use the 'Condition' action to check if the confidence score is below the threshold.. To handle low-confidence extractions, you need to evaluate the confidence score and then take action. The 'Condition' action checks if the score is below the threshold, and the 'Send an email' action notifies a reviewer. Together, they implement the required logic. Logging, terminating, or looping do not directly handle the case but may be supplementary.

Answer

Use the 'Create a record' action in Dataverse to log the low-confidence document.

Answer

Use the 'Terminate' action to stop the flow when confidence is low.

Answer

Use the 'Apply to each' action to iterate over all extracted fields.