AB-410 — frequently asked questions

Automated cloud flows run asynchronously after a record event, so they cannot block or immediately warn a user during form entry; they also duplicate logic per flow.

Yes. Table-level business rules apply to all forms for that table and can be scoped as needed, which reduces duplicated validation across forms.

Canvas apps do not render Dataverse model-driven forms, so the rule has no form context to attach to. Use Entity/Table scope so the rule evaluates against the table data.

It limits the rule to that form; if the form is used only in the intended app, the effect is app-only. Other apps or forms using the Account table are unaffected.

A Word template produces a document from data, usually through Power Automate, and is not a generative AI prompt inside the app. The case requires reusable generative AI with dynamic inputs and minimal custom code.

Yes. Prompts can be invoked from Power Apps, Power Automate, and Copilot Studio, matching Fabrikam's requirement for reusable generative AI solutions.

A canvas app is a user interface; the scenario needs autonomous background analysis, orchestration, and storage, which are handled by AI Builder, cloud flow, and Dataverse.

No. A business rule enforces logic on Dataverse records but cannot retrieve external feedback, call an AI prompt, or orchestrate an escalation workflow.

A prompt column is bound to one table, so it cannot act as a shared instruction asset for forms on several tables. An AI Builder custom prompt is reusable at the environment level.

Yes. The Prompt control can reference a reusable AI Builder custom prompt, allowing the same tone, structure, and required-field instructions to run across forms for multiple tables.

It supports only Approve and Reject, so it cannot represent the Escalate to manager outcome; use Custom Responses - Wait for one response instead.

Yes. With Custom Responses - Wait for one response, the approval completes on the first decision and does not wait for the remaining approvers.

Moderation filters harmful or disallowed content categories; it does not check whether a claim is grounded in your internal support knowledge, so fabricated facts can still pass.

No. It makes answers more consistent and conservative while still responding; only very low values risk terse or repetitive output, so test a moderate low setting.

Do until re-evaluates its condition each iteration, so it can keep running until the status field equals Complete; Apply to each only walks a collection that already exists.

Yes. Apply to each iterates the array retrieved at the start, so it does not pick up records added later and processes each item in that fixed collection once.

The Customer column is a polymorphic lookup to an Account or Contact record and stores no tier value. A Choice column holds the fixed tier options such as Gold, Silver and Bronze.

An alternate key creates a unique index over one or more columns, so external systems can look up records by a stable business value instead of the GUID and duplicates are rejected.

Display order only changes where BPFA appears in the business process flow switcher and which flow is the default; users can still switch to BPFA if it remains active and they have privileges.

No. Dataverse business rules control field behavior and values, not the list of available business process flows, so they cannot enforce the 'must not use BPFA' requirement.

Cloud flows run asynchronously on the server after a Dataverse event, so nothing is pushed to the open form; the user would need to refresh to see any change.

Use JavaScript form event handlers (OnLoad, OnChange, OnSave) with formContext controls, or business rules for simpler visibility and requirement logic.

Identity fields such as the username/UPN are synced from Microsoft Entra ID and read-only in Dataverse, so the field cannot be changed in the app.

Yes. Dataverse links the systemuser record by object ID, so updating the UPN in Entra ID keeps the same user record and its app history.

The mailbox record only stores email server and synchronization settings. Authentication uses the UPN on the Microsoft Entra ID user object, which the mailbox rename never updates.

No. The Dataverse user record linked to the app history stays intact; updating the UPN in Microsoft Entra ID renames the identity without recreating the user.

Production is a role/status. Managed is the technical environment type that enforces the required governance policies.

Managed environments enforce solution layering and prevent unauthorized changes, ensuring governance compliance.

Assign the users or their team a Dataverse security role that has the privileges the app requires, typically through the app's sharing settings. A license alone does not grant access.

It reports whether the user is licensed, which security roles and business unit apply to them, and the specific missing privilege causing the access-denied error.

Add Find Columns only decides which columns the Quick Find search box scans. The field remains in the Advanced Find filter picker because its Searchable property is still enabled.

No. Searchable only affects search and Advanced Find metadata exposure. The field still displays on the form and remains visible to users with appropriate privileges.

Entity scope runs the rule on all forms, views, and the Web API, while a form scope runs only on the selected form—matching the requirement that logic run against the Account Main form only.

The Recommendation action: it shows a recommendation message and a suggested action on the form when the trigger condition (for example, a drop-down value) is met.

It is row-scoped, so it fires on every insert, update, or delete across the table, including unrelated record changes, which the company explicitly wants to avoid.

Yes. The 'When an action is performed' trigger can be filtered to the named action, so only SubmitForReview invocations start the flow.

No, Business Process Flows are for guiding users through stages. To send emails, you must create a separate Cloud Flow (Power Automate) triggered by the record creation.

BPF steps are available in the Model-Driven App. Offline capability depends on the app's offline settings, but the BPF itself does not manage offline sync; the app does.

No. Records already in the flow continue with the version they started on; the new stage applies only to records created after the new version is published.

You cannot force an in-progress instance onto the new version through publication alone; the existing process instances keep their original version until they are completed or the records are re-initialized.

Power Automate's SharePoint connector does not create list items through an atomic changeset, so several independent item creations cannot be rolled back as one transaction; a compensating Scope is required instead.

No. A Scope only groups the actions so failures can be detected; you still need a second step whose Run After is set to 'Has failed' to run the cleanup that deletes or corrects the created items.

To guide users through a set of defined stages, ensuring they complete necessary actions in a logical sequence.

Use a BPF for multi-stage, guided workflows in Model-Driven Apps. Use buttons for specific, isolated actions within Canvas Apps.

Power Fx calls services through predefined connectors with managed sign-in; it cannot build raw HTTP requests with custom auth schemes or custom response parsing as this scenario requires.

Choose a plugin for event-driven, transactional logic and outbound calls to external web services; dataflows are for scheduled, batch data movement with fixed connectors.

A PCF component is custom code and therefore not native consumption. The model already exposes a Power Fx function, so custom components add needless complexity.

No. A stored prediction returns an old result; the requirement is to generate new predictions in real time while the user interacts with the app.

User2's access is defined by direct reports, not by position level. Manager hierarchy at depth 1 limits her to direct reports' data; position hierarchy with depth 2 would grant indirect-report access and break least privilege.

The business unit scopes User1 to finance-only data, and the manager field points to User2, which is the link manager hierarchy uses. Without both, the hierarchy has nothing to grant and User1's data remains out of scope.

A custom page is a separate canvas surface that must be built and embedded; it does not automatically render alongside the model-driven view of rows. A system chart already provides the parallel chart pane.

A chart is created against an entity and paired with a view, so it displays next to whichever view is active. Managers can switch views while the chart updates to reflect the filtered rows.

Yes. Client-side scripts attached to form events call the control's setVisible method to hide or show columns immediately without a page refresh or server round trip.

Yes—show/hide can also be done with business rules or modern form logic. The question only asks whether the JavaScript solution meets the goal, and multiple solutions may be correct in this set.

The checker resolves access for a single user principal; a Dataverse team is a permission container and cannot be evaluated as the subject whose access is in question.

The Testers team is an Entra group team tied to Contoso Ltd. QA, so members added to that Microsoft Entra security group inherit the team and its Quality Assurance role once group membership replicates.

A named formula only defines one app-scoped value or auto-calculating expression, so it cannot be parameterized or shared between apps the way a user-defined function plus a component library can.

Functions are defined per app, but a component library distributes the reusable UI and its behavior, which is how the shared implementation reaches every canvas app.

The case explicitly requires minimizing custom AI development. A pre-built model combined with a prompt template can generate reports aligned to an existing template without building a custom model from scratch.

A blank prompt increases development effort and inconsistency, which the case already flags as a problem. A prompt template provides a standardized starting point with dynamic input support for reuse across services.

Yes. A business rule can set a single-select Choice (option set) column to a specific value or clear it. Only multi-select Choices columns and a few specialized column types are unsupported.

Solution export and import carries schema and customizations such as tables, columns and site map changes, but it never carries records. Data rows require the Configuration Migration tool.

It reversibly decompresses a Dataverse solution file into multiple XML files for source-control and re-compresses them, so it supports ALM workflows rather than transporting components between environments.

The table on the 'many' side. For Vendor to Shipment, the Shipment table receives the Vendor lookup column, making Vendor the parent (1) side.

Cascade All deletes every related child record regardless of state or ownership. Cascade Active would delete only related records that are active.

Input variables are the prompt action's declared parameters; referencing them in the prompt text lets user-supplied values flow into the model's instructions.

Conversation variables carry values captured earlier in the conversation (from agent or bot responses) so a prompt action can read them without asking the user again.

A retry policy automatically re-attempts a failed connector action after a delay, so brief downstream service outages recover without manual intervention instead of failing the whole run.

No. Increasing the timeout only extends how long Power Automate waits for one response; it does not re-send a request that already failed, so retries are required for transient errors.

Prompt inputs such as the customer message must be defined and passed to the prompt before it runs; otherwise the prompt has no data to evaluate and returns no meaningful analysis.

Use 'Set output in variable' immediately after 'Execute prompt' so the prompt's response is captured in a flow variable that subsequent actions can reference.

Cloud flows are queued and processed asynchronously after the Dataverse operation commits, so they cannot execute within the same transaction or roll back with it.

Low-code plugins provide the same synchronous, in-pipeline execution as classic plugins but are authored with minimal code and configured in the Power Apps maker portal, avoiding C# project setup.

Change tracking only flags which rows changed for integration tools like dataflows or Synapse Link; it stores no old/new column values, so it cannot maintain a historical log.

Yes. Column-level auditing is ignored unless auditing is enabled for the table (and at the organization level), so the Account table switch must be turned on first.

Use List rows on the Qualification table, filtered to the rows related to the current Service Request, so every qualification can be listed in the message.

The template must still be populated per qualification as Valid or Not Valid, so the flow has to loop over retrieved rows and compose the body before sending.

Because only the parent Service Request was reassigned; without a cascading assign behavior, the child Qualification records keep the original owner's access and remain hidden from the substitute.

A 1:N (one-to-many) relationship, with Service Request as the parent and Qualification as the child, since one Service Request can have many Qualification records.

A connection is environment- and user-bound, so it cannot be imported with a solution. The connection reference is the portable object that lets IT's service accounts rebind the flow in each environment.

No. A business process flow guides a user through stages; it cannot programmatically set the status field or send the client results email when all qualifications finish.

Because a service request can have one or more Qualification records, Apply to each iterates them so every qualification can be listed in the client email body.

The Condition tests each qualification's verification outcome so the flow can append either Valid or Not Valid for that specific record.

Rollup columns only aggregate related records using count, sum, min, max, or average — they cannot generate a random number, and the sampling must target Qualification rows, not service request headers.

Sequential autonumbers distribute evenly across the table, so exactly one in every ten values is divisible by 10, producing a consistent 10 percent sample of Qualification rows for QV double-checking.

Security roles apply environment-wide to every record of a table, so a role cannot limit a record to its one assigned verifier. Access teams grant record-level permissions that can be added and removed per Service Request.

Because access is granted per record, the substitute is simply added to the access team for the affected Service Requests and gains visibility of all related Qualification records without broadening anyone else's access.

Adoption Fee is a currency column, so Power Fx formula columns need an explicit cast. Decimal('Adoption Fee') * 0.20 guarantees a correctly typed decimal result.

In Power Fx, 20% is treated as the number 20, so 'Adoption Fee' * 20% returns 2,000% of the fee. Use 0.20 to represent twenty percent.

Weight is a field on the Dog table, but staff work from the resident record. A form component bound to the dog lookup surfaces that field inline so they can edit it in context.

Select the lookup that relates the resident record to the dog — the Dog table lookup — so the embedded form shows the correct dog's data.

The weight field is stored as a column on the Dog table, not the Resident table. Updating it from a Resident form still requires Write privileges on Dog in the security role.

The case requires staff to view adopter information for upcoming meetings but explicitly prohibits them from updating adopter data, so Write on Contact is excluded.

A business rule applies conditional logic and can show a validation error, but it cannot display a visible, ordered multi-stage list of duties on the screen.

That is a separate validation concern, typically handled by a business rule or column validation, while the ordered duty list itself is shown by the business process flow.

A business rule runs on the client while the date is entered, so it can display the error immediately and block an invalid Pickup date. A flow runs asynchronously and would allow the bad date to be saved first.

Use DateDiff('Spay and neuter date','Pickup date'), which counts days from the spay date to the pickup date, and require the result to be at least 2 days so a two-day pickup is accepted.

Run after settings only define behavior when the previous action fails, is skipped, or times out. They do not pause a flow; the start-and-wait approval action is what suspends the flow until a decision is submitted.

No. Secure inputs only mask data in run history for privacy and have no effect on approver routing, waiting for a decision, or branching on the approval outcome.

The Lock/Unlock action, available in the Actions group of the business rule visual designer, makes a column read-only when the rule's conditions are satisfied.

A Recommendation analyzes column values and suggests next best actions with supporting reasons to help users complete the form accurately.

Connections backed by Microsoft Entra ID can expire after long periods without use, so actions relying on that existing connection configuration can no longer authenticate to Dataverse. Reauthenticating the connection restores it.

Insert a condition or input validation step before the failing action to check for null or missing values, then branch incomplete records to an error path instead of letting the action run and fail.

Retraining is a model-improvement step, not an inference call, and doing it per submission would be slow and costly while still not providing a native in-app prediction mechanism.

Exporting moves the model between environments; it does not let the running app call the model. Native consumption requires invoking the model at runtime, for example with Power Fx.

The rule is most likely still inactive. You must activate the process after adding the Set Default value action for it to run when the form loads.

No. A form-load default belongs in a business rule evaluated on the form; a cloud flow runs asynchronously and will not reliably set the value as the form opens.

No, Choice controls are used in Canvas Apps forms. Business Process Flows use specific step types like Condition, Stage, and Power Automate Cloud Flow.

You must insert a 'Power Automate Cloud Flow' step into the BFF sequence. This step allows you to select an existing flow to run before the user proceeds to the next stage.

The Predict() function is used to execute a prompt model and return its result within a canvas app.

You must add the AI Builder data connection to your app so that the custom prompt appears in the list of available data sources.

App Assistant Agents provide contextual, in-app guidance to users using Copilot. Agent Flows are background processes that automate business logic triggered by specific events.

No, Agent Flows are designed for backend execution and automation. Contextual user guidance is the specific role of App Assistant Agents.

Managed solutions are imported as read-only to prevent unauthorized changes. You must develop in an unmanaged solution to allow edits.

It allows users to modify components, leading to instability and making upgrades difficult. Production should always use managed solutions.

Add Existing includes the selected component and its current configuration. Add Required automatically adds any components that the selected component depends on.

By carefully selecting only the specific table or column using Add Existing, you avoid pulling in unrelated dependencies that might come with Add Required.

Adding images to the solution ensures they are version-controlled and deployed correctly. Assigning them without this step can cause missing resource references.

You typically save individual table configurations and then perform a single 'Publish all customizations' action at the end to make all changes live simultaneously.

Canvas apps allow for highly customizable user interfaces tailored specifically for capturing unique input fields, whereas Model-Driven apps follow a rigid structure based on the data model.

Yes, AI Builder actions are available as triggers or steps within Cloud flows to process data, but the core 'autonomous evaluation' capability belongs to the AI Builder service itself.

Because the Error() call runs inside the same formula that IFERROR wraps, IFERROR catches it and returns its fallback value instead of surfacing an error.

No. Power Fx error handling is scoped to the single formula that performs the wrap; other formulas must handle their own errors.

A semantic model is a Power BI data artifact that defines measures and relationships; it renders nothing inside the app. Dashboards provide the actual in-app visual display next to rows.

Choose a custom page when you need bespoke, canvas-style layouts or logic not covered by native dashboard components. For standard charts beside a row list, a dashboard is simpler and native.

A global variable created with Set is available on every screen for the life of the app session, while a context variable passed via Navigate exists only on the destination screen until another navigation occurs.

A multicolumn collection stores full records with multiple fields per row (e.g., customer ID, name, and amount), which is needed for tabular data. A single-column collection stores only a list of individual values.

The app picks the highest form in the form order that the user's security role has access to. Roles with no access to higher forms fall back to the next available one.

No. Form availability, form order, and role assignments are solution customizations, so you must publish the customizations before app users see the change.

Open the form, use Assign security roles, and select only the roles that should see it; users without an assigned role cannot open that form for the table.

Yes. In the view designer you apply security roles to a system view, and it disappears from the Views list for users who hold none of those roles.

A PCF component is custom code you author yourself. Native consumption means using the built-in AI Builder integration that Power Fx calls directly, with no custom component required.

Stored predictions are results from an earlier run on different data. The requirement is to generate predictions during user interaction, so the model must be invoked live for the current image.

Table-level permissions control the whole table, so they either leave the view exposed to everyone or block non-managers from the entire table instead of just this one view.

No. Hiding navigation only removes the shortcut; users can still reach the view and see its sensitive columns, so it provides no real access control.

Auditing writes a retained history of create, update, and delete actions for the table's columns, while change tracking flags modified rows so integrations can pull only the changed data. They solve two different requirements.

No. Change tracking only indicates which rows have changed since a point in time for synchronization; the durable history of who changed what is stored by auditing in the audit log.

You can embed Power BI reports, but system charts are the native, lightweight solution for displaying simple aggregated metrics directly on the entity form.

System charts typically appear at the top of the main form view, allowing users to see trends and aggregates while editing or viewing a single record.

No. Components within a Managed solution are read-only in the target environment. You must export it as a Managed solution to preserve these restrictions.

The Default solution is typically unmanaged and editable. It does not provide the immutability required to stop display name changes in downstream environments.

In the maker portal, open Data > Tables > Contacts > Columns, select the column, and set Field security to Yes. The column then becomes editable only through a column security profile.

Grant the profile Read on the secured column, and leave Create, Update (Write), and Delete unselected so users can see but not modify the value.

Set the column's requirement level to Business required on the form or column, so Dataverse blocks the save until the column is populated.

Use a business rule with a Show/hide column action, evaluating your condition so the extra columns appear only when it is true.

Owner teams exist per business unit by default, so team-based roles grant access automatically to all members and survive user changes without reassignment.

Organization-owned tables do appear in security roles, but their records carry no owning business unit, so the business unit requirement fails.

No. Dataverse supports only Organization, User or Team, and None ownership; business units group users, teams, and records rather than owning tables.

A subgrid displays a view of many related records and is not an embedded editable form for a single row. Editing the record in place requires a Form component.

Use a Quick view form when you need read-only fields from one related-table record shown alongside the current record. Use a subgrid when you need a list of related rows.

The question states the framework is already enabled, so it is a prerequisite already satisfied rather than the further action needed to distribute reusable components across apps.

A component library is added to canvas apps as a data source and its republished updates reach all consumers automatically, whereas code component packages are environment-level PCF imports that must be redeployed separately.

Column security profiles restrict access to individual columns (fields), making them read-only or masked. They do not limit which customer rows a user can retrieve, so they cannot deliver row-level scoping.

View filters narrow the rows a specific view returns, while assigning security roles to the view decides which user groups may use that view at all. Both are needed for a complete restriction.

The data step retrieves the department manager's address from the request record, so the email action needs it as a dynamic recipient value and cannot run first.

A trigger fires when a new maintenance request is created in Power Apps, and that trigger must be the first component of the flow.

A custom page is a separate surface that opens apart from the view, so it doesn't render in parallel with the list of rows. A system chart displays directly in the chart pane next to the active view.

A semantic model only defines the underlying data structure for reporting; it produces no visual by itself. The question needs an in-app rendered visual, which a system chart provides natively.

Go to the view settings and change the audience from 'Everyone' to 'Specific security roles', then select the desired roles.

No, table-level permissions control record read/write access. View visibility is controlled separately by assigning the view to security roles.

Separate solutions in one development environment still share the environment's unmanaged layer, dependencies, and connections. The teams can still interfere with each other, so the scenario's overwrite and independent-update requirements call for separate development environments.

Managed solutions support controlled layering, upgrades, and removal in production. They prevent unmanaged customization drift and make releases more predictable than importing unmanaged solutions or editing production directly.

It can work when teams have a shared release cadence and truly disjoint components, with strong solution segmentation and source control. It is weaker when teams must independently update overlapping features or avoid overwriting each other's changes.

Smart assist is a Dynamics 365 Customer Service workspace feature for case suggestions, not a general Power Platform component you add to any model-driven app; Microsoft has also announced its end of support.

Yes. When embedded in a model-driven app, the agent can be given context from the open form or view so its responses reflect the current record and data without the user leaving the app.

It is the in-app AI assistant for model-driven apps, answering user questions from app data and metadata in an assistant pane so users stay within forms and views.

Business Unit access limits visibility to records owned within the manager's business unit. The requirement says all opportunities within their organization, so Organization-level Read matches the stated scope.

No. User-level access already lets representatives read and update records they own. Organization-level access would expose every opportunity, violating the confidential-data restriction.

No. Read and Write are separate privileges. Organization-level Read gives visibility across all opportunities without granting update rights unless Write is also configured at that level.

Setting Searchable to No keeps that column out of Advanced Find and related filter column choices, so users cannot select the non-address1 address columns when building views. The column and its data remain in the table.

Column-level security controls whether users can read or edit data in a secured column. It does not replace the Searchable setting for removing a column from Advanced Find/view filter selection, and removing read access from all users is too broad.

No. Business rules primarily affect form behavior and field visibility on forms; they do not control the column list used to create views and filters in Dataverse.

A native Dataverse many-to-many link records only the association, not effective dates, roles, or assignment context. Connections (or a custom intersect table with history fields) preserve each assignment as a separate record.

Create a self-referential one-to-many relationship on the Customer table using a lookup such as Parent Customer, then enable the hierarchical setting. Each child customer points to one parent, and a parent can have many children.

It is the same relationship viewed from opposite ends: create a one-to-many from Customer to Support Case, which appears as a many-to-one lookup on the case. Choose the one-to-many option when that is the available answer.

Records remain linked to you because Dataverse uses your immutable Entra Object ID (GUID) for ownership, not your email address.

You can change your display name in Outlook or Teams, but to change your sign-in name (UPN), you must use the Microsoft 365 admin center.

Plug-ins run on the Dataverse server. The case explicitly wants server-side custom code minimized, so a client-side JavaScript web resource is the better fit for form extensions.

PCF controls package custom UI as reusable components that can be used across model-driven and canvas apps, reducing the duplication described in the case.

A plug-in would fit if the requirement demanded server-side logic that must run regardless of the client, such as enforcing a process on every create or update. That is not the stated extensibility requirement here.

The requirement says the name is persisted on the home screen only. A context variable is created on and readable from that screen, while Set() creates a global variable visible to every screen and control, which is wider scope than required.

No. A gallery needs a table of records, so the work orders must be materialised into a collection with ClearCollect or Collect. A named formula is a read-only computed value, not a row set.

The requirements call for responsive design, full control over the UI, and taking photos on mobile devices. Canvas apps provide pixel-level layout control and the Camera and Add picture controls, which model-driven forms do not.

Solution checker validates the solution, but it is only a quality gate. The update still must be exported and deployed to production as a managed solution to provide controlled release and removal.

Yes. Deleting a managed solution removes its components from the target environment, subject to dependencies and shared components. That clean removal behavior is why managed solutions are preferred for production releases.

Unmanaged solutions are used for development and remain editable. Managed solutions are packaged for deployment, cannot be directly modified in the target environment, and provide better release control for test and production.

App designer controls which tables are included in the app, but the site map defines areas, groups and subareas. Without site map customization, added tables do not become business-process groupings.

Use the model-driven app's navigation and site map editing experience so changes apply to that app. Editing a shared default solution site map can affect other apps that use it.

No. Adding only required tables scopes what the app exposes; security roles and business unit membership still govern record-level access.

A stacked column places sales stages in one column per region, so the column height shows total pipeline and each segment shows that stage's contribution. Clustered columns compare stage values side by side without the regional total.

It removes the runtime chart picker from the dashboard component, so users cannot switch that component to another chart. The configured chart and default view remain fixed.

Apply it as the dashboard component's default View. That view controls the records loaded into the chart at runtime, separately from the chart's category and stacked-series aggregation.

Data retrieval in App.OnStart blocks the app's rendering. Removing it allows the UI to load instantly while data fetches continue in the background, improving perceived performance.

It executes multiple data source calls simultaneously in parallel, reducing the total wait time compared to running them one after another.

App visibility determines if an app appears in the App Launcher for a user. Security roles define what data and actions a user can perform within that app. An app can be visible but inaccessible if the user lacks the required security role.

No. In Dataverse, table permissions (privileges) are defined within security roles. You must create or modify a security role to include the necessary privileges, and then assign that role to the user.

Calling the prompt describes where it is used, not how it is configured to accept dynamic inputs. The requirement focuses on making the prompt adaptable, which requires input variables.

Input variables allow you to pass dynamic data into a prompt template at runtime, enabling the same prompt to generate different outputs based on varying inputs across different scenarios.

A Custom Connector is a component that allows you to connect your Power Platform apps and flows to any REST API, including those requiring authentication and custom request handling.

No. Environment variables can only store static values like URLs or keys. You still need a connector or flow action to actually make the HTTP request.

No. Formula columns support arithmetic and string manipulation but cannot call external AI models. You must use the Prompt data type for AI-generated text.

Business Rules automate simple UI or data changes based on conditions. Prompt columns use AI Builder to generate new content (like summaries) based on input data fields.

Context provides the data source (e.g., PDF, SharePoint list) that the AI reads. Expectations define the rules for the output, such as what topics to cover, the tone, or specific details to include.

Creating a Word document is an output action or format, not a content expectation. The prompt's expectation section should define the textual content of the summary, not the container it ends up in.

Variables store dynamic data values (like user names) to be inserted into text. Templates provide the complete, structured instruction set (system prompt) that can be reused and customized across different topics without rewriting the core logic.

Yes. When you add a template from the library to a topic, you can modify its instructions, add specific context, or adjust parameters while keeping the original template intact for other uses.

Yes, but Business Rules apply to the entire entity form. If the requirement is strictly for the BPF stage progression, configuring the Data Step as Required is more precise and adheres to the scope constraint.

You can configure a Button action in the BPF stage that triggers a Power Automate cloud flow, or use a Record-based trigger in the flow if the data update coincides with the stage change.

No, Business Rules operate locally on the current entity's form context. For cross-entity validation, you must use a Plugin, JavaScript, or a Flow.

The system will prevent the record from being saved if the field is empty when the condition is met, displaying a standard error message to the user.

Yes, but it introduces latency. For immediate button-click responses where data is already on the screen, calling the prompt directly in app logic is faster and simpler.

You reference the text box value (e.g., TextInput1.Text) as an argument or parameter in the AI.Prompt function within your app code.

A 401 error occurs when the credentials for a connector have expired, been revoked, or are invalid, requiring the user to reauthenticate.

Concurrency control limits how many instances of an action can run at the same time, preventing redundant processing of the same trigger event.

The syntax is greater(value1, value2), where value1 is the variable being checked and value2 is the threshold number.

In standard Power Automate expressions, contains() is generally case-insensitive, meaning 'vip' and 'VIP' will both match.

No. BPF triggers fire when a user moves to a new stage or performs an action within the flow, not when underlying data fields are updated programmatically or manually outside the flow context.

It fires when the row is modified, regardless of whether the changed field was read-only. However, you can filter the trigger in Power Automate to only run when specific columns are updated.

Variables must be defined with a specific name and type before they can be used. Initializing creates the container in the flow's memory.

Yes. You must Initialize a variable before you can Set it. Setting a non-existent variable causes an error.

Classify assigns a category label to a document based on its content. Process documents extracts specific key-value pairs or table data from the document structure.

Yes, the 'Process documents' action utilizes Azure AI Document Intelligence capabilities under the hood. You must have access to these services within your Power Platform environment.

A calculated column evaluates its expression on access using columns from the same row (plus limited lookups). A rollup column aggregates related rows and is stored, refreshed periodically or on demand by the scheduled rollup job.

Both persist or recompute a value rather than computing it at read time. A business rule writes to the row on save, and JavaScript only runs in the client form context — neither satisfies "not stored" and the JavaScript option also violates the custom-code constraint.

Calculated columns are available for retrieval and can be used in many queries, but heavy use of calculated columns across large tables can hurt query performance, so reserve them for row-level derived values like this scenario.

You choose an authentication type when creating the connector — API key, OAuth 2.0, basic authentication, or a no-auth/custom scheme — and the platform injects the credentials on every call. Keys and secrets are stored securely rather than hardcoded in flows.

Use a custom connector when the target REST API has no built-in connector in Power Platform. You supply the OpenAPI definition describing the endpoints, request and response bodies, and the authentication scheme.

Yes. After the custom connector is published in an environment, its operations appear in cloud-flow action pickers, so a Dataverse row-created or row-updated trigger can call the external AI service and store the returned data in Dataverse.

The requirement tracks changes to the Technician lookup column on Maintenance Requests, not edits to technician master data. Auditing the Technicians table would create different audit records and would not capture previous and current lookup values on the maintenance request.

The environment and table may generate audit records for other audited items, but the specific lookup change history will not be captured. You must enable column auditing on the Technician column to get previous and current technician values with timestamps.

No; it changes the scope. You still need global, table, and column auditing, but you should avoid auditing the whole Technicians table or unrelated columns.

If a column is empty, accessing its value usually returns null or undefined. Comparing null to a string like 'High' using equality ('==') results in false, preventing the flow from executing.

No, typically you need separate triggers for 'When an item is created' and 'When an item is modified'. A single trigger instance cannot listen to two different event types simultaneously.

If the column doesn't exist or isn't populated, the expression triggerBody()?['Priority']?['Value'] will likely resolve to null, causing the equality check against 'High' to fail.

No. A license allows you to run the app, but you still need a Dataverse security role to view or edit data.

While Microsoft Entra groups often sync to Dataverse teams, there can be delays. Direct role assignment ensures immediate access.

Technically, Dataverse defines relationships as One-to-Many. Many-to-One is simply the view of that same relationship from the 'many' side. You cannot create a standalone Many-to-One relationship; it is always part of a One-to-Many pair.

Use Many-to-Many when records in both tables can be associated with multiple records in the other table, such as Students enrolling in Courses. This creates an implicit intersection table behind the scenes.

Add the flow through Power Apps, then set the button's OnSelect to flow.Run(...) and pass any required parameters. This fires the flow only when the button is selected.

No. SubmitForm() submits the form's data to its data source; it does not invoke a Power Automate flow. Use the flow's Run() method from OnSelect after validating or submitting the form.

OnVisible runs when the screen is displayed, not when the user selects Submit, so it can send the email too early or repeatedly on navigation.

Bar Charts compare discrete categories. Line Charts are preferred for time-series data because they emphasize the continuity and direction of the trend over months.

Yes, functionally they are similar for showing proportions. However, exam questions typically specify Pie Charts for simple status breakdowns unless 'donut' is explicitly mentioned as an option.

Even with equal weight, a Funnel Chart is still appropriate for showing 'progression' through stages, as it highlights the volume remaining at each step of the pipeline.

No. SendEmailToUsers sends emails to users' email clients. For mobile push notifications delivered to the Power Apps mobile app, you must use the Power Apps Notification connector and the SendPushNotification() function.

Not necessarily. If the notification is triggered directly by a button click in a canvas app, you can use the Power Apps Notification connector's client-side functions. Cloud Flows are better suited for background automation or when the trigger is a data change.

Match returns only the first matching reference, while MatchAll returns every match in the connector response. Concat can then combine the Value column into the single string the question requires.

Yes. Contains allows the pattern to match anywhere inside the alphanumeric text; BeginsWith or EndsWith would fail because the alphabetic prefix and suffix vary.

Use the exact-eight-digit pattern because the reference format is fixed. A generic digit pattern may capture the wrong sequence or a fragment, especially if other numeric runs appear.

A Canvas page allows you to embed a canvas app within a model-driven app, providing flexibility to connect to various data sources outside of Dataverse without writing custom code.

Only if the third-party data is imported into Dataverse as a custom entity. Otherwise, you need an intermediate layer like a Canvas app or API connector.

No. The Admin Center is for environment management. You must use the Power Apps maker portal to create and export solutions.

The Maker Portal is for building and packaging apps/flows. The Admin Center is for administrators to manage environments, users, and system configurations.

Columns are attributes of the table. To add them, you must edit the table component itself in the solution, as columns do not exist as independent updatable entities outside their parent table.

Change tracking logs when records are created, updated, or deleted. Enabling it at the table level allows auditing these changes, which is critical for compliance and synchronization scenarios.

In the Power Platform admin center or maker portal, when publishing or updating a solution, you can choose to publish only specific selected components rather than the entire solution to minimize impact.

Choice columns store values as integers in the database, whereas Text columns store the actual string characters. Integers occupy significantly less space than variable-length strings.

No. Text columns are limited to single lines. You must use Multiline Text to preserve line breaks and support longer content, though it consumes more storage.

Unmanaged solutions allow full modification of components in target environments, which is essential for iterative development and testing of schema changes like new columns.

A publisher provides a unique prefix for solution components, ensuring they are clearly identified and do not conflict with system or third-party components.

Developer environments are intended for individual learning and prototyping. They do not support the security group management or formal ALM processes required for enterprise team development.

ALM (Application Lifecycle Management) requires a structured flow from development (Sandbox) to production (Production). This ensures changes are tested and controlled before affecting business operations.

No, only users with appropriate permissions in Power Apps Studio can use the Monitor tool to connect to and analyze other users' sessions.

The session data will be lost when the user disconnects or closes the app, making it impossible to perform the required 'later review' of the issue.

A Formula Column calculates values on-the-fly and does not store them in the database. A Business Rule sets a field's value, which is then stored in the database.

Plugins are custom code solutions that typically update field values, causing storage. They also violate the 'minimize custom code' requirement compared to native Formula Columns.