Scenario practice questions

You are creating a business process flow in Dataverse. What is the primary purpose of this component?

Trap 1: To act as a chatbot

Business process flows guide users through defined stages of a business process, enforcing consistent data entry across model-driven apps. Chatbot capability belongs to Microsoft Copilot Studio, which would be the right tool when conversational interaction is the requirement rather than staged process guidance.

Trap 2: To perform complex calculations

Business process flows structure the stages and steps a user follows to complete a process; they do not evaluate expressions or compute values. Complex calculations belong in calculated columns, rollup columns, or Power Automate expressions, which would be correct when derived numeric results are needed.

Trap 3: To automate background tasks

Business process flows are interactive, user-driven stage trackers; they do not execute unattended logic. Background automation is the role of Power Automate cloud flows or workflows, which would be the correct choice when tasks must run without a user actively progressing the record.

A
To guide users through a business process
Business process flows in Dataverse provide a guided, stage-based interface that steers users through predefined steps, enforcing a consistent path to completion. This directly fulfils the stem's requirement by structuring how users progress through a process, ensuring every record follows the same defined stages and required data entry.
B
To act as a chatbot
Why it fails: Business process flows guide users through defined stages of a business process, enforcing consistent data entry across model-driven apps. Chatbot capability belongs to Microsoft Copilot Studio, which would be the right tool when conversational interaction is the requirement rather than staged process guidance.
C
To perform complex calculations
Why it fails: Business process flows structure the stages and steps a user follows to complete a process; they do not evaluate expressions or compute values. Complex calculations belong in calculated columns, rollup columns, or Power Automate expressions, which would be correct when derived numeric results are needed.
D
To automate background tasks
Why it fails: Business process flows are interactive, user-driven stage trackers; they do not execute unattended logic. Background automation is the role of Power Automate cloud flows or workflows, which would be the correct choice when tasks must run without a user actively progressing the record.

Question 2mediummultiple choice

A developer needs to create a relationship between two Dataverse tables: 'Project' and 'Task'. A project can have many tasks, but each task belongs to exactly one project. The developer also wants to ensure that when a project is deleted, all related tasks are automatically deleted. Which type of relationship and configuration should they use?

Trap 1: One-to-many relationship with cascade delete set to 'Remove Link'.

'Remove Link' only removes the reference between the records, leaving the task records orphaned. The requirement is to delete the tasks when the project is deleted, so 'Cascade All' is needed. 'Remove Link' is used when you want to keep the related records but disassociate them.

Trap 2: One-to-many relationship with cascade delete set to 'Restrict'.

'Restrict' prevents the deletion of a project if it has related tasks. This is the opposite of the requirement, which is to automatically delete tasks. Therefore, 'Restrict' is incorrect. 'Cascade All' is the appropriate setting for automatic deletion.

Trap 3: Many-to-many relationship with cascade delete enabled.

Many-to-many relationships allow multiple projects to be associated with multiple tasks. However, the requirement is that each task belongs to exactly one project, which is a one-to-many relationship. Cascade delete is not applicable in many-to-many relationships in the same way; it would not automatically delete tasks when a project is deleted.

A
One-to-many relationship with cascade delete set to 'Remove Link'.
Why it fails: 'Remove Link' only removes the reference between the records, leaving the task records orphaned. The requirement is to delete the tasks when the project is deleted, so 'Cascade All' is needed. 'Remove Link' is used when you want to keep the related records but disassociate them.
B
One-to-many relationship with cascade delete set to 'Restrict'.
Why it fails: 'Restrict' prevents the deletion of a project if it has related tasks. This is the opposite of the requirement, which is to automatically delete tasks. Therefore, 'Restrict' is incorrect. 'Cascade All' is the appropriate setting for automatic deletion.
C
Many-to-many relationship with cascade delete enabled.
Why it fails: Many-to-many relationships allow multiple projects to be associated with multiple tasks. However, the requirement is that each task belongs to exactly one project, which is a one-to-many relationship. Cascade delete is not applicable in many-to-many relationships in the same way; it would not automatically delete tasks when a project is deleted.
D
One-to-many relationship with cascade delete configured to 'Cascade All'.
A one-to-many relationship allows a single project to be associated with many tasks, and each task to be associated with one project. Configuring cascade delete to 'Cascade All' ensures that when a project is deleted, all related task records are also deleted automatically. This matches the requirement.

Question 3mediummultiple choice

You need to create a field that automatically calculates the total revenue of a customer based on related invoices. Which field type should you use?

Trap 1: Lookup field

A lookup field copies a value from a related record; it cannot aggregate multiple related invoices into a sum. It is tempting because lookups do relate records, and would be right for pulling a single field, such as a customer's credit limit, from one related row.

Trap 2: Calculated field

A calculated field computes values at query time from the same record's columns and cannot aggregate across related invoice rows. It is tempting because it performs arithmetic automatically, but rollup fields are the type designed to sum related records.

Trap 3: Formula column

A formula column computes values at read time from other columns; it cannot aggregate across related invoice rows. It is tempting because formulas handle arithmetic on the same record, which is correct when calculating a value such as line-item total from quantity and price within one row.

A
Lookup field
Why it fails: A lookup field copies a value from a related record; it cannot aggregate multiple related invoices into a sum. It is tempting because lookups do relate records, and would be right for pulling a single field, such as a customer's credit limit, from one related row.
B
Calculated field
Why it fails: A calculated field computes values at query time from the same record's columns and cannot aggregate across related invoice rows. It is tempting because it performs arithmetic automatically, but rollup fields are the type designed to sum related records.
C
Rollup field
A rollup field aggregates values from related child records, so it can sum invoice amounts linked to a customer and display the total automatically. It recalculates as related invoices change, satisfying the stem's requirement for an automatically calculated total revenue value.
D
Formula column
Why it fails: A formula column computes values at read time from other columns; it cannot aggregate across related invoice rows. It is tempting because formulas handle arithmetic on the same record, which is correct when calculating a value such as line-item total from quantity and price within one row.

Question 4easymultiple choice

A developer is building a canvas app that must let a user pick a single date from a calendar-style picker. The selected date then filters a gallery of records. Which control should the developer add to the screen for this purpose?

Trap 1: Dropdown

A Dropdown lists predefined choices and does not present a calendar for date selection. While it could be populated with dates, it is not designed for date picking and lacks the SelectedDate property, making it a poor fit for filtering by a chosen calendar date in this scenario.

Trap 2: Slider

A Slider produces a numeric value within a range and has no concept of dates or calendars. Using it to represent a date would require mapping numbers to dates manually, which is awkward and does not deliver a date picker experience, so it does not meet the requirement.

Trap 3: Text input

A Text input accepts free-form text, so the user could type a date, but it offers no calendar, no validation, and the value is a string that must be parsed before filtering. This adds error-prone conversion logic and fails to provide the calendar-style picker the scenario requires.

A
Dropdown
Why it fails: A Dropdown lists predefined choices and does not present a calendar for date selection. While it could be populated with dates, it is not designed for date picking and lacks the SelectedDate property, making it a poor fit for filtering by a chosen calendar date in this scenario.
B
Slider
Why it fails: A Slider produces a numeric value within a range and has no concept of dates or calendars. Using it to represent a date would require mapping numbers to dates manually, which is awkward and does not deliver a date picker experience, so it does not meet the requirement.
C
Date picker
The Date picker control provides a calendar interface for selecting one date, and its SelectedDate property can be referenced in the gallery's Items filter. This directly satisfies the requirement of picking a single date that drives a filter, without extra configuration for multi-date or time selection.
D
Text input
Why it fails: A Text input accepts free-form text, so the user could type a date, but it offers no calendar, no validation, and the value is a string that must be parsed before filtering. This adds error-prone conversion logic and fails to provide the calendar-style picker the scenario requires.

Question 5easymultiple choice

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?

Trap 1: When an item is created

The 'When an item is created' trigger is for SharePoint lists, not document libraries. While a document library is technically a list, this trigger is designed for list items and may not provide the file-specific metadata needed to retrieve and process the file content with AI Builder. It is not the intended trigger for file-based scenarios.

Trap 2: When a HTTP request is received

This trigger starts the flow when an HTTP POST is sent to a generated URL. It is not automatically invoked by SharePoint when a file is added. Using it would require an external system to detect the new file and call the flow, adding unnecessary complexity and not meeting the requirement of a native SharePoint trigger.

Trap 3: When a file is modified (properties only)

This trigger fires when an existing file is modified, not when a new file is added. The scenario requires the flow to run on new file creation, so using a modification trigger would miss the initial upload and could cause the flow to run multiple times on subsequent edits, leading to incorrect processing.

A
When an item is created
Why it fails: The 'When an item is created' trigger is for SharePoint lists, not document libraries. While a document library is technically a list, this trigger is designed for list items and may not provide the file-specific metadata needed to retrieve and process the file content with AI Builder. It is not the intended trigger for file-based scenarios.
B
When a file is created (properties only)
The SharePoint trigger 'When a file is created (properties only)' fires when a new file is added to the specified library. It provides file properties, including the file identifier, which the flow can use to retrieve the file content and pass it to AI Builder. This is the correct trigger for reacting to new files in SharePoint.
C
When a HTTP request is received
Why it fails: This trigger starts the flow when an HTTP POST is sent to a generated URL. It is not automatically invoked by SharePoint when a file is added. Using it would require an external system to detect the new file and call the flow, adding unnecessary complexity and not meeting the requirement of a native SharePoint trigger.
D
When a file is modified (properties only)
Why it fails: This trigger fires when an existing file is modified, not when a new file is added. The scenario requires the flow to run on new file creation, so using a modification trigger would miss the initial upload and could cause the flow to run multiple times on subsequent edits, leading to incorrect processing.

Question 6mediummultiple choice

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

Trap 1: Many-to-One

Many-to-One links many child rows to a single parent, so each Contact could reference only one Account. It is tempting because it models a lookup where many records share one owner, which is correct when each Contact belongs to exactly one Account rather than several.

Trap 2: One-to-Many

One-to-Many gives one Account many Contacts, but each Contact still points to a single Account, so the reverse many-to-many pairing fails. It is tempting because it is the standard Account-to-Contact lookup, correct when Contacts belong to one Account only.

Trap 3: Parental

Parental is a cascade-behaviour variant of One-to-Many, not a cardinality, so it still restricts each Contact to one Account. It is tempting because parental relationships enforce cascade delete and assign ownership, which is correct when child records must inherit the parent's owner.

A
Many-to-One
Why it fails: Many-to-One links many child rows to a single parent, so each Contact could reference only one Account. It is tempting because it models a lookup where many records share one owner, which is correct when each Contact belongs to exactly one Account rather than several.
B
Many-to-Many
A many-to-many relationship uses an intersect table to link records in both directions, so each contact can relate to several accounts and each account to several contacts. One-to-many would restrict a contact to a single account, failing the scenario.
C
One-to-Many
Why it fails: One-to-Many gives one Account many Contacts, but each Contact still points to a single Account, so the reverse many-to-many pairing fails. It is tempting because it is the standard Account-to-Contact lookup, correct when Contacts belong to one Account only.
D
Parental
Why it fails: Parental is a cascade-behaviour variant of One-to-Many, not a cardinality, so it still restricts each Contact to one Account. It is tempting because parental relationships enforce cascade delete and assign ownership, which is correct when child records must inherit the parent's owner.

Question 7mediummultiple choice

You are designing a Dataverse solution for a real estate agency. The agency needs to track properties and the agents who list them. Each property can have only one listing agent, but an agent can list many properties. You create a table named Property and a table named Agent. You need to establish the relationship so that when an agent is deleted, their listed properties remain but the Listing Agent column is cleared. Which relationship type and behavior should you configure?

Trap 1: One-to-many relationship with Referential, Restrict Delete behavior…

This relationship type fits the cardinality, but Referential, Restrict Delete prevents deleting an agent who has related properties. The requirement is to allow agent deletion and clear the lookup, so this behavior would cause an error instead of the desired outcome.

Trap 2: One-to-many relationship with Parental behavior on the Agent side.

Parental behavior would cascade the delete to related Property records, deleting them when the Agent is removed. The requirement is to keep properties and only clear the Listing Agent column, so Parental would cause unintended data loss and is incorrect.

Trap 3: Many-to-many relationship with Referential, Restrict Delete…

Many-to-many would allow multiple agents per property, contradicting the requirement that each property has only one listing agent. Also, Referential, Restrict Delete would block deletion of an agent if any property references them, which is not desired because properties should remain with a cleared lookup.

A
One-to-many relationship with Referential, Remove Link behavior on the Agent side.
A one-to-many relationship from Agent to Property matches the business rule that one agent lists many properties. Referential, Remove Link sets the Listing Agent lookup on related Property records to null when the Agent is deleted, preserving the Property records. This is exactly the required deletion behavior.
B
One-to-many relationship with Referential, Restrict Delete behavior on the Agent side.
Why it fails: This relationship type fits the cardinality, but Referential, Restrict Delete prevents deleting an agent who has related properties. The requirement is to allow agent deletion and clear the lookup, so this behavior would cause an error instead of the desired outcome.
C
One-to-many relationship with Parental behavior on the Agent side.
Why it fails: Parental behavior would cascade the delete to related Property records, deleting them when the Agent is removed. The requirement is to keep properties and only clear the Listing Agent column, so Parental would cause unintended data loss and is incorrect.
D
Many-to-many relationship with Referential, Restrict Delete behavior.
Why it fails: Many-to-many would allow multiple agents per property, contradicting the requirement that each property has only one listing agent. Also, Referential, Restrict Delete would block deletion of an agent if any property references them, which is not desired because properties should remain with a cleared lookup.

Question 8mediummultiple choice

You need to automate document extraction using AI Builder. The documents are varied invoices. Which model type should you configure?

Trap 1: Object Detection

Object Detection locates and classifies visual objects within images, not text fields on invoice pages. It is tempting because invoices are scanned images, and object detection would be correct for tasks such as identifying damage or products in photographs.

Trap 2: Entity Extraction

Entity Extraction targets prebuilt entities in structured forms, not the varied field layouts of invoices; it cannot infer vendor-specific totals or dates. It suits custom scenarios with consistent, labelled fields. Invoice processing needs the prebuilt Invoice model, which handles diverse invoice formats and extracts key fields automatically.

Trap 3: Prediction model

Prediction models classify or forecast from structured or tabular data, not extract fields from document layouts. They are tempting because invoices feed downstream forecasting, and a prediction model would be correct where historical data must predict outcomes such as late payment risk.

A
Document Processing
Document Processing is the AI Builder model type designed for extracting structured fields such as totals, dates and vendor names from varied invoice layouts. It satisfies the varied invoices constraint by learning from labelled samples rather than requiring a fixed template.
B
Object Detection
Why it fails: Object Detection locates and classifies visual objects within images, not text fields on invoice pages. It is tempting because invoices are scanned images, and object detection would be correct for tasks such as identifying damage or products in photographs.
C
Entity Extraction
Why it fails: Entity Extraction targets prebuilt entities in structured forms, not the varied field layouts of invoices; it cannot infer vendor-specific totals or dates. It suits custom scenarios with consistent, labelled fields. Invoice processing needs the prebuilt Invoice model, which handles diverse invoice formats and extracts key fields automatically.
D
Prediction model
Why it fails: Prediction models classify or forecast from structured or tabular data, not extract fields from document layouts. They are tempting because invoices feed downstream forecasting, and a prediction model would be correct where historical data must predict outcomes such as late payment risk.

Question 9mediummultiple choice

You want to use AI Builder to predict whether a customer will churn. You have a Dataverse table with historical churn data. What is the first step?

Trap 1: Create an object detection model.

Object detection classifies and locates items within images, which churn prediction from tabular Dataverse records does not involve. It is tempting because it is a genuine AI Builder model type, and it would be correct for scenarios such as reading product images or shelf photos, but churn requires a prediction model trained on historical outcome data.

Trap 2: Export the data to Excel.

AI Builder trains directly on Dataverse tables, so exporting to Excel breaks the native data connection and adds no required preparation step. It is tempting because Excel is familiar for cleaning and exploring data, and it would be reasonable for offline analysis, but the first step here is selecting the Dataverse table and its historical churn columns.

Trap 3: Train a model in the AI Hub.

Training cannot begin before a model type is chosen and its training data selected; the AI Hub is where you build after that preparation. It is tempting because training is the visible core activity, and it would be correct once the prediction model and labelled Dataverse data are configured, but it is not the first step.

A
Create an object detection model.
Why it fails: Object detection classifies and locates items within images, which churn prediction from tabular Dataverse records does not involve. It is tempting because it is a genuine AI Builder model type, and it would be correct for scenarios such as reading product images or shelf photos, but churn requires a prediction model trained on historical outcome data.
B
Export the data to Excel.
Why it fails: AI Builder trains directly on Dataverse tables, so exporting to Excel breaks the native data connection and adds no required preparation step. It is tempting because Excel is familiar for cleaning and exploring data, and it would be reasonable for offline analysis, but the first step here is selecting the Dataverse table and its historical churn columns.
C
Train a model in the AI Hub.
Why it fails: Training cannot begin before a model type is chosen and its training data selected; the AI Hub is where you build after that preparation. It is tempting because training is the visible core activity, and it would be correct once the prediction model and labelled Dataverse data are configured, but it is not the first step.
D
Create a Prediction model in AI Builder.
AI Builder prediction models require you to first create the model, then select the Dataverse table and target column before training. Creating the Prediction model initiates the workflow that consumes the historical churn data as training input.

Question 10easymulti select

Which THREE of the following are benefits of using Copilot in Power Automate? (Choose THREE)

Trap 1: Replacing the need for Dataverse

Copilot in Power Automate generates flows and expressions from natural language; it does not replace Dataverse, which remains the data platform for storing and modelling business data. It is tempting because Copilot reduces manual authoring effort, which can feel like removing the need for underlying data services.

Trap 2: Automatically fixing all syntax errors in expressions

Copilot in Power Automate suggests expressions and can explain or repair them, but it does not automatically fix all syntax errors; validation still flags malformed expressions for the maker to correct. It is tempting because Copilot genuinely assists with authoring and debugging expressions, which is where syntax errors commonly surface.

A
Explaining complex flow logic
Copilot in Power Automate can describe what a flow does in plain language, clarifying triggers, actions and expressions. This helps makers understand inherited or complex flows, satisfying the benefit of explaining logic without manually tracing every step.
B
Replacing the need for Dataverse
Why it fails: Copilot in Power Automate generates flows and expressions from natural language; it does not replace Dataverse, which remains the data platform for storing and modelling business data. It is tempting because Copilot reduces manual authoring effort, which can feel like removing the need for underlying data services.
C
Automatically fixing all syntax errors in expressions
Why it fails: Copilot in Power Automate suggests expressions and can explain or repair them, but it does not automatically fix all syntax errors; validation still flags malformed expressions for the maker to correct. It is tempting because Copilot genuinely assists with authoring and debugging expressions, which is where syntax errors commonly surface.
D
Troubleshooting flow run failures
Copilot in Power Automate analyses failed runs and surfaces likely causes with suggested fixes, reducing diagnostic effort. This directly addresses troubleshooting flow run failures, letting makers resolve errors without manually inspecting each action's inputs and outputs.
E
Generating flow suggestions from prompts
Generating flow suggestions from prompts lets makers describe automation intent in natural language and receive a working cloud flow scaffold, removing the need to know triggers and actions upfront. This directly satisfies the stem's benefit criterion by accelerating flow creation for users unfamiliar with Power Automate's connector model.

Question 11mediummultiple choice

You need to allow a user to select multiple items from a list in a canvas app. Which control should you use?

Trap 1: Radio button

A Radio button control enforces a single mutually exclusive selection within a group, so it cannot capture multiple items. It is tempting because radio buttons present list options clearly, and they are the correct choice when exactly one option must be chosen from several.

Trap 2: Toggle

A Toggle returns a single Boolean value, so it cannot hold multiple selections from a list. It is tempting because toggles suit binary on/off choices such as enabling a feature, which is the scenario where a Toggle would be the correct control.

Trap 3: Dropdown

A Dropdown control allows only one selected value at a time, so it cannot return multiple items. It is tempting because dropdowns handle long lists compactly, and they are correct when a user must pick a single item from many options.

A
Radio button
Why it fails: A Radio button control enforces a single mutually exclusive selection within a group, so it cannot capture multiple items. It is tempting because radio buttons present list options clearly, and they are the correct choice when exactly one option must be chosen from several.
B
Combo box
A combo box supports multi-select through its SelectMultiple property, storing chosen items as a table the app can bind and iterate. Dropdowns and list boxes either restrict selection to one value or lack the searchable, tag-style selection behaviour required here.
C
Toggle
Why it fails: A Toggle returns a single Boolean value, so it cannot hold multiple selections from a list. It is tempting because toggles suit binary on/off choices such as enabling a feature, which is the scenario where a Toggle would be the correct control.
D
Dropdown
Why it fails: A Dropdown control allows only one selected value at a time, so it cannot return multiple items. It is tempting because dropdowns handle long lists compactly, and they are correct when a user must pick a single item from many options.

Question 12easymultiple choice

You need to create a column that stores the total cost of a project, which is derived from the sum of all associated expense records. Which column type should you use?

Trap 1: Lookup column

A Lookup column retrieves a single related record's value from another table; it cannot aggregate multiple related expense rows into a sum. It is tempting because it does relate tables, but it is designed for one-to-many single-value retrieval, whereas a Rollup column performs the required SUM aggregation.

Trap 2: Calculated column

A calculated column computes from columns in the same row, so it cannot aggregate related expense records in another table. It is the right choice for row-level arithmetic such as line totals, whereas summing child records requires a rollup column.

Trap 3: Choice column

A Choice column stores one value from a predefined option set, so it cannot hold a numeric sum. It is correct for capturing categorised selections such as status or priority, not for deriving totals from related expense records.

A
Lookup column
Why it fails: A Lookup column retrieves a single related record's value from another table; it cannot aggregate multiple related expense rows into a sum. It is tempting because it does relate tables, but it is designed for one-to-many single-value retrieval, whereas a Rollup column performs the required SUM aggregation.
B
Calculated column
Why it fails: A calculated column computes from columns in the same row, so it cannot aggregate related expense records in another table. It is the right choice for row-level arithmetic such as line totals, whereas summing child records requires a rollup column.
C
Choice column
Why it fails: A Choice column stores one value from a predefined option set, so it cannot hold a numeric sum. It is correct for capturing categorised selections such as status or priority, not for deriving totals from related expense records.
D
Rollup column
A rollup column aggregates values from related records in a one-to-many relationship, applying functions such as SUM across child expense rows. It stores the computed total cost on the parent project record, matching the stem's requirement for a derived sum of associated expenses.

Question 13easymultiple choice

You are creating a new custom table in Dataverse to store customer feedback. You need to ensure that each feedback record is associated with a specific account. Which type of column should you add to the feedback table to create this association?

Trap 1: A currency column.

A currency column stores monetary values and has no capability to reference another table. It is used for financial data such as amounts and prices. It cannot create a relationship or store a reference to an account. Therefore, it does not fulfill the requirement of associating a feedback record with a specific account.

Trap 2: A choice column.

A choice column provides a fixed list of options defined within the column. It cannot reference records in another table. While you could list account names as choices, this would be static and not maintain a relationship. It would not allow you to access account details or enforce that the selected value corresponds to an existing account record. It is not suitable for creating a dynamic association.

Trap 3: A text column.

A text column stores free-form text and cannot create a relationship to another table. While you could store the account name as text, it would not enforce referential integrity or allow you to easily navigate to the account record. It also would not support features like lookups or related views. This does not meet the requirement for associating a feedback record with an account.

A
A currency column.
Why it fails: A currency column stores monetary values and has no capability to reference another table. It is used for financial data such as amounts and prices. It cannot create a relationship or store a reference to an account. Therefore, it does not fulfill the requirement of associating a feedback record with a specific account.
B
A choice column.
Why it fails: A choice column provides a fixed list of options defined within the column. It cannot reference records in another table. While you could list account names as choices, this would be static and not maintain a relationship. It would not allow you to access account details or enforce that the selected value corresponds to an existing account record. It is not suitable for creating a dynamic association.
C
A text column.
Why it fails: A text column stores free-form text and cannot create a relationship to another table. While you could store the account name as text, it would not enforce referential integrity or allow you to easily navigate to the account record. It also would not support features like lookups or related views. This does not meet the requirement for associating a feedback record with an account.
D
A lookup column.
A lookup column creates a many-to-one relationship from the feedback table to the account table. It allows selecting a single account record and stores a reference to that record. This enforces referential integrity and enables navigation to the account from the feedback record. It is the standard way to associate a record with another table in Dataverse.

Question 14hardmulti select

You are creating a Copilot in Copilot Studio. Which TWO features help ensure that the Copilot provides accurate and relevant answers?

Trap 1: Increasing the model temperature to 1.0.

Temperature 1.0 maximises sampling randomness, producing varied and less deterministic answers, which undermines accuracy and relevance. It tempts because higher temperature suits creative brainstorming or marketing copy generation, where diverse phrasing matters more than factual grounding.

Trap 2: Disabling authentication.

Authentication governs access control, not answer grounding, so disabling it cannot improve accuracy and merely exposes the Copilot to anonymous users. It is tempting because removing sign-in friction feels like it broadens data reach, yet authentication is genuinely required when the Copilot must restrict responses to a signed-in user's own Microsoft Entra ID-scoped content.

Trap 3: Using only custom-coded plugins.

Custom-coded plugins only extend actions and data retrieval; they supply no grounding mechanism, so accuracy still depends on generative answers configuration. Plugins are the right choice when the Copilot must call an external API or automate a task, not when the requirement is ensuring responses draw on authoritative knowledge sources.

A
Increasing the model temperature to 1.0.
Why it fails: Temperature 1.0 maximises sampling randomness, producing varied and less deterministic answers, which undermines accuracy and relevance. It tempts because higher temperature suits creative brainstorming or marketing copy generation, where diverse phrasing matters more than factual grounding.
B
Adding Knowledge sources (e.g., SharePoint, Website).
Knowledge sources ground the Copilot in authoritative organisational content such as SharePoint sites and websites, so answers derive from verified material rather than model guesswork. This directly satisfies the requirement for accurate, relevant responses within the defined scope.
C
Disabling authentication.
Why it fails: Authentication governs access control, not answer grounding, so disabling it cannot improve accuracy and merely exposes the Copilot to anonymous users. It is tempting because removing sign-in friction feels like it broadens data reach, yet authentication is genuinely required when the Copilot must restrict responses to a signed-in user's own Microsoft Entra ID-scoped content.
D
Using only custom-coded plugins.
Why it fails: Custom-coded plugins only extend actions and data retrieval; they supply no grounding mechanism, so accuracy still depends on generative answers configuration. Plugins are the right choice when the Copilot must call an external API or automate a task, not when the requirement is ensuring responses draw on authoritative knowledge sources.
E
Defining clear System instructions (System prompt).
System instructions constrain the Copilot's behaviour, tone and scope before any user turn, directly shaping how it interprets requests and which knowledge it draws on. For accurate, relevant answers, this prompt-level grounding steers generative responses away from off-topic output, satisfying the requirement without relying on external data connectors or post-response filtering.

Question 15mediummultiple choice

You are building a Power Automate cloud flow that must call a Copilot Studio agent to summarize a long customer email. The email body is 15,000 characters. The agent is configured with a custom prompt and expects a single string input. Which action should you use to invoke the agent and pass the email body?

Trap 1: Use the 'Invoke an HTTP endpoint' action with the agent's Direct…

While Direct Line can be used to talk to bots, it is not the recommended or simplest way to invoke a Copilot Studio agent from Power Automate. It requires managing secrets, constructing activity payloads, and handling conversation state manually. The 'Run a Copilot' action abstracts all of this and is the correct choice for passing a single input string.

Trap 2: Use the 'Send an HTTP request to SharePoint' action to post the…

This action targets SharePoint REST endpoints, not Copilot Studio agents. It cannot authenticate to or address a Copilot Studio agent's runtime endpoint, and it would not surface the agent's declared input parameters. Using it here would fail at configuration time because no agent endpoint is exposed through that connector. It is the wrong connector for agent invocation.

Trap 3: Use the 'Create a chat session' action in the Azure AI Bot Service…

The Azure AI Bot Service connector is for Azure Bot Service resources, not Copilot Studio agents surfaced through the Power Platform connector. It would require additional channel registration and would not directly map the agent's input parameter. For Copilot Studio agents, the supported flow action is the Copilot connector's 'Run a Copilot' action, not a bot session creation action.

A
Use the 'Invoke an HTTP endpoint' action with the agent's Direct Line secret and pass the email body as an activity.
Why it fails: While Direct Line can be used to talk to bots, it is not the recommended or simplest way to invoke a Copilot Studio agent from Power Automate. It requires managing secrets, constructing activity payloads, and handling conversation state manually. The 'Run a Copilot' action abstracts all of this and is the correct choice for passing a single input string.
B
Use the 'Send an HTTP request to SharePoint' action to post the email body to the agent endpoint.
Why it fails: This action targets SharePoint REST endpoints, not Copilot Studio agents. It cannot authenticate to or address a Copilot Studio agent's runtime endpoint, and it would not surface the agent's declared input parameters. Using it here would fail at configuration time because no agent endpoint is exposed through that connector. It is the wrong connector for agent invocation.
C
Use the 'Create a chat session' action in the Azure AI Bot Service connector and pass the email body as a message.
Why it fails: The Azure AI Bot Service connector is for Azure Bot Service resources, not Copilot Studio agents surfaced through the Power Platform connector. It would require additional channel registration and would not directly map the agent's input parameter. For Copilot Studio agents, the supported flow action is the Copilot connector's 'Run a Copilot' action, not a bot session creation action.
D
Use the 'Run a Copilot' action and map the email body to the agent's input parameter.
The 'Run a Copilot' action in Power Automate is the supported connector action for invoking a Copilot Studio agent from a cloud flow. It exposes the agent's declared input parameters, so mapping the email body string to the agent's input parameter is correct. It also returns the agent's response for downstream steps. This is the standard integration path for agent invocation from flows.

Question 16easymultiple choice

You are building a canvas app and need to show the signed-in user a temporary success banner at the top of the screen after they submit a form. The banner should appear immediately and disappear on its own after a few seconds without navigating away. Which Power Apps feature should you use?

Trap 1: A Timer control that resets the form after the interval elapses.

A Timer can drive timed behavior, but resetting the form is unrelated to showing a success banner and could discard the user's submitted data. Using a Timer to reveal and hide a Label is possible but far more complex than the built-in transient notification the scenario calls for.

Trap 2: A Navigate call to a dedicated success screen.

Navigating to another screen changes context and does not produce a temporary banner that auto-dismisses; the user would have to navigate back. The scenario explicitly says the message should appear and disappear without navigating away, so screen navigation is the wrong mechanism.

Trap 3: A Label control whose Visible property is set to true permanently.

A Label toggled always visible would remain on screen indefinitely, not disappear after a few seconds. It also occupies layout space and requires manual show/hide logic, so it does not satisfy the temporary, self-dismissing banner behavior described in the scenario.

A
A Timer control that resets the form after the interval elapses.
Why it fails: A Timer can drive timed behavior, but resetting the form is unrelated to showing a success banner and could discard the user's submitted data. Using a Timer to reveal and hide a Label is possible but far more complex than the built-in transient notification the scenario calls for.
B
A Navigate call to a dedicated success screen.
Why it fails: Navigating to another screen changes context and does not produce a temporary banner that auto-dismisses; the user would have to navigate back. The scenario explicitly says the message should appear and disappear without navigating away, so screen navigation is the wrong mechanism.
C
A Label control whose Visible property is set to true permanently.
Why it fails: A Label toggled always visible would remain on screen indefinitely, not disappear after a few seconds. It also occupies layout space and requires manual show/hide logic, so it does not satisfy the temporary, self-dismissing banner behavior described in the scenario.
D
A Notify function call with the NotificationType and a timeout.
Notify displays a transient banner at the top of the app and automatically dismisses it after the specified duration, which matches the requirement for a temporary success message that disappears without navigation. It is designed exactly for short-lived user feedback and requires no extra control or timer logic.

Question 17mediummultiple choice

You are building a model-driven app for a healthcare clinic. The clinic wants to track patients and their appointments. A patient can have many appointments, and each appointment is associated with exactly one patient. When a patient is deleted, all their appointments should also be deleted. Which relationship behavior should you configure between the Patient and Appointment tables?

Trap 1: Referential, Restrict Delete

Referential, Restrict Delete prevents deletion of a patient if they have related appointments. The requirement is to delete appointments when the patient is deleted, so this behavior would block the deletion and leave orphaned records, which is not desired.

Trap 2: Referential, Remove Link

Referential, Remove Link would set the patient lookup on appointments to null when the patient is deleted. However, the requirement is to delete the appointments entirely, not just clear the link. This behavior would leave appointment records without a patient, which is not acceptable.

Trap 3: Configurable Cascading with Cascade Active

Configurable Cascading with Cascade Active would delete only active appointments, leaving inactive ones. The requirement is to delete all appointments regardless of status. This behavior is more granular and does not meet the need for unconditional deletion of all related appointments.

A
Parental
Parental behavior creates a cascade delete: when the parent (Patient) is deleted, all related child records (Appointments) are also deleted. This matches the requirement that appointments should be removed when a patient is deleted. It also enforces that an appointment cannot exist without a patient.
B
Referential, Restrict Delete
Why it fails: Referential, Restrict Delete prevents deletion of a patient if they have related appointments. The requirement is to delete appointments when the patient is deleted, so this behavior would block the deletion and leave orphaned records, which is not desired.
C
Referential, Remove Link
Why it fails: Referential, Remove Link would set the patient lookup on appointments to null when the patient is deleted. However, the requirement is to delete the appointments entirely, not just clear the link. This behavior would leave appointment records without a patient, which is not acceptable.
D
Configurable Cascading with Cascade Active
Why it fails: Configurable Cascading with Cascade Active would delete only active appointments, leaving inactive ones. The requirement is to delete all appointments regardless of status. This behavior is more granular and does not meet the need for unconditional deletion of all related appointments.

Question 18hardmultiple choice

You are a solution architect for a large enterprise. The company uses Dataverse and needs to ensure that when a custom table named Invoice is created or updated, a series of complex calculations are performed and the results are stored in related records across multiple tables. The calculations involve aggregating data from several other tables and must run reliably even if initiated from bulk data imports or background processes. You need to minimize development effort and ensure the logic is reusable across multiple apps. What should you use?

Trap 1: A business rule on the Invoice table with actions to set values on…

Business rules cannot update records in other tables. They are limited to the table they are defined on and can only set values, show/hide columns, enable/disable columns, and lock columns on the current record. They cannot perform aggregations across multiple tables or update related records, making them unsuitable for this complex scenario.

Trap 2: A classic workflow that runs on create and update of the Invoice…

Classic workflows are deprecated and run asynchronously, which may not provide the immediate consistency required. They also have limited capabilities for complex calculations and aggregations. While they can update related records, they are not recommended for new development and may not run during bulk imports reliably. Plug-ins are the modern, supported approach.

Trap 3: A Power Automate cloud flow triggered on create and update of the…

Power Automate cloud flows can trigger on Dataverse row changes, but they are asynchronous and may not run immediately, especially during bulk imports. They also have limits on the number of actions and can be less reliable for complex, high-volume calculations. While reusable, they may not meet the reliability and performance requirements for complex aggregations across multiple tables.

A
A business rule on the Invoice table with actions to set values on related tables.
Why it fails: Business rules cannot update records in other tables. They are limited to the table they are defined on and can only set values, show/hide columns, enable/disable columns, and lock columns on the current record. They cannot perform aggregations across multiple tables or update related records, making them unsuitable for this complex scenario.
B
A classic workflow that runs on create and update of the Invoice table.
Why it fails: Classic workflows are deprecated and run asynchronously, which may not provide the immediate consistency required. They also have limited capabilities for complex calculations and aggregations. While they can update related records, they are not recommended for new development and may not run during bulk imports reliably. Plug-ins are the modern, supported approach.
C
A Power Automate cloud flow triggered on create and update of the Invoice table.
Why it fails: Power Automate cloud flows can trigger on Dataverse row changes, but they are asynchronous and may not run immediately, especially during bulk imports. They also have limits on the number of actions and can be less reliable for complex, high-volume calculations. While reusable, they may not meet the reliability and performance requirements for complex aggregations across multiple tables.
D
A Dataverse plug-in registered on the Create and Update messages of the Invoice table.
Dataverse plug-ins run synchronously or asynchronously within the platform's event pipeline, ensuring they execute reliably even during bulk imports and background processes. They can perform complex calculations, aggregate data from multiple tables, and update related records. Plug-ins are reusable across apps and provide the most control and reliability for server-side business logic.

Question 19mediummultiple choice

You are an app builder for a logistics company. The company uses a Dataverse table named Shipment with a choice column named ShipmentStatus that includes options such as Draft, InTransit, Delivered, and Cancelled. Business rules must automatically set the ShipmentStatus to Delivered when the ActualDeliveryDate column is populated, but users should not be able to manually change the status to Delivered if the ActualDeliveryDate is empty. Which Dataverse feature should you use to implement this requirement?

Trap 1: Create a real-time workflow that triggers on create and update of…

Real-time workflows run server-side and can enforce logic, but they do not lock columns in the user interface. Users could still attempt to set Delivered manually; the workflow would only block it after submission with an error, providing a poor user experience. The requirement is to prevent manual changes, not just reject them after the fact.

Trap 2: Create a column-level security profile that makes the…

Column-level security can make a column read-only, but it applies to all users and cannot conditionally allow edits based on other column values. Additionally, Power Automate cloud flows are asynchronous by default and do not provide immediate UI feedback. This approach would prevent legitimate status updates and lacks the conditional logic needed.

Trap 3: Write a JavaScript web resource that registers an OnChange event…

JavaScript web resources work only in model-driven apps and require custom code, which increases maintenance. They do not automatically apply to canvas apps or other clients. The requirement can be met with declarative business rules that work across clients without code, making JavaScript unnecessarily complex and limited in scope.

A
Create a real-time workflow that triggers on create and update of Shipment, sets ShipmentStatus to Delivered when ActualDeliveryDate is not null, and throws an error if a user attempts to set Delivered without ActualDeliveryDate.
Why it fails: Real-time workflows run server-side and can enforce logic, but they do not lock columns in the user interface. Users could still attempt to set Delivered manually; the workflow would only block it after submission with an error, providing a poor user experience. The requirement is to prevent manual changes, not just reject them after the fact.
B
Create a business rule that sets the ShipmentStatus column to Delivered when ActualDeliveryDate is not null and locks the ShipmentStatus column when ActualDeliveryDate is null.
Business rules run client-side and server-side to enforce logic without code. Setting the status when ActualDeliveryDate is populated and locking the column when it is empty ensures users cannot manually set Delivered prematurely. This matches the requirement exactly because business rules can both set values and lock columns based on conditions.
C
Create a column-level security profile that makes the ShipmentStatus column read-only for all users and use a Power Automate cloud flow to update it when ActualDeliveryDate changes.
Why it fails: Column-level security can make a column read-only, but it applies to all users and cannot conditionally allow edits based on other column values. Additionally, Power Automate cloud flows are asynchronous by default and do not provide immediate UI feedback. This approach would prevent legitimate status updates and lacks the conditional logic needed.
D
Write a JavaScript web resource that registers an OnChange event handler on the ActualDeliveryDate column to set ShipmentStatus and an OnSave event handler to validate that Delivered is only set when ActualDeliveryDate is populated.
Why it fails: JavaScript web resources work only in model-driven apps and require custom code, which increases maintenance. They do not automatically apply to canvas apps or other clients. The requirement can be met with declarative business rules that work across clients without code, making JavaScript unnecessarily complex and limited in scope.