AB-410 Practice Exam 3 — 20 Questions

Question 1

You are building a canvas app that uses an AI Builder model to classify customer messages. You need the app to display the classification result immediately after the user submits a message. Which approach should you use?

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

Accepted Answer

Call the AI Builder model directly from the canvas app using the AI Builder connector and bind the result to a label.. The AI Builder connector in Power Apps is designed for direct, in-app model invocation. Binding the prediction to a label gives the user immediate feedback. The other options introduce delays or use features that cannot run AI Builder models in the required context.

Answer

Use a Power BI tile to visualize the model output and embed it in the canvas app.

Answer

Create a Power Automate flow that runs the model and writes the result to a SharePoint list, then refresh the app's data source.

Answer

Configure a Dataverse business rule to call the AI Builder model when a new row is created.

Question 2

You are configuring an AI Builder custom document processing model to extract line items from invoices. The model performs well on vendor name and invoice date but consistently misses the line item table. You have already uploaded 50 labeled samples. What should you do first to improve table extraction?

Accepted Answer

Add more labeled samples that include a variety of line item table layouts and formats.. Poor table extraction usually stems from insufficient variety in labeled examples. Adding samples that show different line item table layouts gives the model the patterns it needs to generalize. Switching model type, increasing iterations without new data, or changing the document type does not address the root cause and risks harming the fields that already perform well.

Answer

Increase the number of training iterations to 100 without changing the dataset.

Answer

Switch the model to a prebuilt invoice model and remove the custom labels.

Answer

Change the document type to 'Receipt' in the model configuration.

Question 3

You are building a Power Automate cloud flow that must create a new record in Microsoft Dataverse and then immediately send a Microsoft Teams message containing the record's unique identifier. The flow currently uses the 'Add a new row' action followed by 'Post message in a chat or channel'. During testing, the Teams message sometimes posts before the Dataverse record is fully committed, so the identifier is missing. What should you do to ensure the Teams message always includes the correct identifier?

Accepted Answer

Use the 'Add a new row' action's output directly in the Teams message and remove any parallel branches.. The Dataverse 'Add a new row' action returns the new record's unique identifier as an output. To ensure the Teams message includes it, the Teams action must run after the Dataverse action completes and must reference that output. Removing parallel execution and using the action's dynamic content directly enforces the correct order and data dependency.

Answer

Replace the Dataverse action with an HTTP request to the Dataverse Web API and set 'Retry Policy' to none.

Answer

Add a 'Delay' action of 10 seconds between the Dataverse action and the Teams action.

Answer

Enable 'Secure Inputs' and 'Secure Outputs' on the Dataverse action.

Question 4

A Power Automate cloud flow uses a Copilot Studio agent to classify support tickets. The agent returns a JSON object with a 'category' field. You need to branch the flow based on the category value and then log the ticket ID and category to a Microsoft Dataverse table. Which combination of actions should you use to parse the agent response and write the record?

Accepted Answer

Use 'Parse JSON' with a schema generated from the agent response, then use 'Add a new row' in the Microsoft Dataverse connector.. Agent responses that are JSON objects must be parsed before individual fields can be used in expressions or branches. 'Parse JSON' with a schema makes the 'category' field available as dynamic content, enabling a Switch or Condition. The Microsoft Dataverse 'Add a new row' action then persists the ticket ID and category to the target table, satisfying both the branching and storage requirements.

Answer

Use 'Compose' to store the raw response, then use 'Create record' in the Common Data Service (current environment) connector.

Answer

Use 'Initialize variable' to hold the response, then use 'Insert row' in the SQL Server connector.

Answer

Use 'Select' to map the response fields, then use 'Create item' in the SharePoint connector.

Question 5

You are designing a Power Automate cloud flow that processes incoming invoices. The flow uses AI Builder to extract key fields, then routes the invoice for approval if the total exceeds $10,000. You need to ensure that the approval request includes the extracted invoice number and total. Which configuration should you use?

Accepted Answer

Add a 'Start and wait for an approval' action and include dynamic content from the AI Builder output in the approval's details field.. The 'Start and wait for an approval' action is designed to handle approval workflows and allows dynamic content from previous steps, including AI Builder outputs, to be embedded in the approval details. This ensures approvers see the invoice number and total directly, enabling informed decisions. It also pauses the flow until a response is received, which is essential for conditional routing based on approval outcome.

Answer

Create a SharePoint list to store the extracted fields, then use a 'Start and wait for an approval' action that references the list item ID.

Answer

Configure a 'Condition' action that checks the total, then use a 'Post adaptive card and wait for a response' action in Microsoft Teams with manually typed field names.

Answer

Use a 'Send an email' action with the extracted fields and include an approval link generated by a separate 'Create an approval' action.

Question 6

You have a Power Automate cloud flow that processes invoices using AI Builder. The flow is triggered when a new file is added to a SharePoint document library. You need to ensure that the flow only processes PDF files and extracts key fields. Which action should you add first in the flow?

Accepted Answer

Condition action to check if the file extension is .pdf. To ensure only PDF files are processed, the flow must include a condition that checks the file extension before any AI Builder action. This prevents errors and saves AI Builder credits. The Condition action is the correct first step, as it gates the rest of the flow based on the file type, aligning with best practices for robust automation.

Answer

Get file content using path action

Answer

Parse JSON action to read file metadata

Answer

AI Builder 'Extract information from invoices' action

Question 7

You are creating a Power Automate flow that uses the AI Builder 'Create text with GPT' action to generate a summary. The flow must ensure that the AI-generated content is not stored in the run history for compliance reasons. What should you do?

Accepted Answer

Enable 'Secure Outputs' on the AI Builder action.. Enabling Secure Outputs on the AI Builder action ensures that the generated text is not stored in the flow run history, which is essential for compliance when handling sensitive or regulated content. This setting redacts the output from logs while allowing the flow to use it.

Answer

Disable run history for the entire flow.

Answer

Use a 'Compose' action to store the output and then delete it.

Answer

Set the flow's retention policy to 0 days.

Question 8

You want to automate the process of sending an email whenever a specific field value changes. What is the best way to achieve this?

Accepted Answer

Power Automate flow. Power Automate flows are the best tool for automating an email notification triggered by a field value change in Dataverse because they provide a no-code/low-code designer with a native Dataverse trigger ('When a row is added, modified, or deleted') and built-in email actions (Outlook, Office 365 Outlook). This makes the solution maintainable by business users and avoids custom code deployment.

Answer

Business Rule

Answer

Plugin

Answer

Workflow

Question 9

You are creating a Power Automate cloud flow that must run whenever a new file is added to a specific SharePoint document library. The flow should then send the file to an AI Builder model for processing. Which trigger should you use?

Accepted Answer

When a file is created (properties only). The correct trigger is the SharePoint 'When a file is created (properties only)' trigger, because it is specifically designed to start a flow when a new file is added to a document library. It provides the file properties needed to retrieve the file and pass it to AI Builder for processing, matching the scenario's requirements.

Answer

When an item is created

Answer

When a HTTP request is received

Answer

When a file is modified (properties only)

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 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.

Question 12

You are building a canvas app that displays a list of products from a Dataverse table. You want to show the total price of all products in a label. Which function should you use in the label's Text property?

Accepted Answer

Sum(Products, Price). To display the total price of all products, the label must aggregate the Price column across all records. The Sum function performs this addition. Other aggregate functions like Average, CountRows, and Max return different metrics and would not produce the total price.

Answer

Average(Products, Price)

Answer

Max(Products, Price)

Answer

CountRows(Products)

Question 13

You are creating a model-driven app for a property management company. The app must include a custom page that displays a map of all properties using a third-party JavaScript mapping library, and this page must appear in the main site map alongside standard table views. Which approach should you use to add this custom page?

Accepted Answer

Create a canvas app and embed it as a custom page in the model-driven app, then add the page to the app's navigation.. Custom pages allow you to embed a canvas app inside a model-driven app, giving you the flexibility of canvas controls while keeping the model-driven navigation and security model. You can add the custom page to the app's site map, making it accessible alongside table views. This is the recommended approach for custom UI such as a map.

Answer

Build a separate Power Apps portal and link to it from the model-driven app's navigation using a URL.

Answer

Add a web resource of type HTML to a table form and configure the form as the app's default page.

Answer

Create a Power BI report with a map visual and embed it in the model-driven app using a dashboard.

Question 14

You need to ensure that an AI model only processes documents from a specific language. Where do you validate this?

Accepted Answer

In a flow Condition action. Language validation for an AI model's input is enforced at runtime in the flow, typically with a Condition action that checks the detected or declared language before the document is passed to the model. This gates processing so only documents in the target language proceed.

Answer

In the model's training set

Answer

In the AI Hub settings

Answer

In the Copilot manifest

Question 15

You are creating a canvas app that displays a list of products. You want the app to show a warning icon next to any product whose stock quantity is below 10. The gallery already has an Image control. What should you set the Visible property of the Image control to?

Accepted Answer

ThisItem.Stock < 10. Inside a gallery, each control is bound to a row, and ThisItem exposes that row's fields. Using ThisItem.Stock < 10 in the Visible property evaluates per row, so the warning icon appears only for products with stock below 10. Other approaches either evaluate globally or reference the wrong scope.

Answer

First(Products).Stock < 10

Answer

CountRows(Products) < 10

Answer

Stock < 10

Question 16

You are building a canvas app that submits new rows to a Microsoft Dataverse table from a form. Users report that duplicate rows appear when they click Save twice quickly. You must prevent duplicate submissions while preserving the form's normal submit behavior. What should you do?

Accepted Answer

Wrap the SubmitForm call in an If statement that checks a local variable flag, sets the flag before SubmitForm, and resets it in the form's OnSuccess property.. Guarding SubmitForm with a local variable that is set before the call and cleared in OnSuccess prevents the second rapid click from starting another submission. The flag is evaluated synchronously in the button's OnSelect, so the second event sees it as true and skips the submit. Resetting in OnSuccess keeps the button usable for the next record.

Answer

Set the form's OnSuccess property to call ResetForm and set the button's OnSelect to Notify with a success message.

Answer

Set the SubmitForm button's DisplayMode to Disabled when the form's Unsaved property is false.

Answer

Add a unique constraint on the Dataverse table and rely on the error returned by SubmitForm to block the duplicate.

Question 17

Which tool would you use to create a Prompt template that extracts information from a document?

Accepted Answer

AI Builder Prompt builder. AI Builder's Prompt builder is the dedicated tool for creating prompt templates that use natural language instructions to extract, summarize, or generate information from documents and other inputs. It lets you define a prompt, select a model, and test it against sample data, then use the resulting prompt in Power Automate or Power Apps. This is exactly the tool for building a document-information-extraction prompt.

Answer

Power Apps Studio

Answer

Flow designer

Answer

Dataverse form designer

Question 18

Which type of relationship is required to allow multiple 'Contacts' to be associated with multiple 'Accounts'?

Accepted Answer

Many-to-Many. A Many-to-Many (N:N) relationship in Dataverse is designed for scenarios where records on both sides can be associated with multiple records on the other side — exactly the case for multiple Contacts linked to multiple Accounts. Dataverse implements N:N relationships using an intersect table that stores the pairings.

Answer

Many-to-One

Answer

One-to-Many

Answer

Parental

Question 19

You are building a canvas app that uses an AI Builder model to classify customer messages. You need the app to display the classification result immediately after the user submits a message. Which approach should you use?

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

Accepted Answer

Define the agent's output as a fixed set of category values and validate the returned value against that list in the app.. Reliable classification requires constraining the agent to a known category set and validating responses before they drive routing, plus testing with representative and ambiguous emails to confirm outcomes. Together these make the behavior predictable and verifiable, while unconstrained or unvalidated output leaves routing brittle.

Answer

Increase the agent's creativity setting so it can invent new categories when none fit.

Answer

Use the agent's raw text response directly as the routing key in an If expression.

Answer

Hard-code the category list in the routing expression and skip validating the agent's response.

Question 20

You are designing a Dataverse data model for a field service application. A dispatcher must be able to assign a work order to exactly one technician, but a technician can be assigned many work orders over time. The work order table already exists and the technician table already exists. You need to add the relationship that enforces this business rule with the least administrative effort. What should you create?

Accepted Answer

A one-to-many relationship from the technician table to the work order table, with a lookup column on the work order table.. The business rule describes classic one-to-many cardinality: one technician is associated with many work orders, and each work order carries a single technician reference. Modeling it as a one-to-many relationship anchored on the technician table and adding the lookup on the work order table satisfies the rule directly. Many-to-many and reversed or self-referential designs either permit invalid assignments or cannot span the two tables.

Answer

A self-referential relationship on the work order table that links each work order to its assigned technician record.

Answer

A one-to-many relationship from the work order table to the technician table, with a lookup column on the technician table.

Answer

A many-to-many relationship between the technician table and the work order table using an intersect table.