You are optimizing a Dataverse-triggered cloud flow configured to run 'When a row is added, modified or deleted' on the 'Opportunity' table. Currently, the flow triggers hundreds of times a day when sales reps update fields like 'Description' or 'Close Date', consuming API limits. You only want the flow to execute if the 'Status Reason' is changed to 'Won' AND the 'Estimated Revenue' is greater than $50,000. How should you implement this optimization to prevent the flow from triggering unnecessarily?
check_circle
Correct AnswerA
To optimize Dataverse triggers and prevent a flow from consuming API calls or run allocations unnecessarily, you should use the 'Filter rows' property in the trigger settings. By providing an OData filter expression (`statuscode eq 3 and estimatedvalue gt 50000`), the Dataverse event engine evaluates the condition server-side. If the condition is not met, the flow is completely ignored and does not trigger at all. Using a 'Condition' action inside the flow (Option 1) means the flow still triggers and consumes resources before terminating.
Question 112
A custom 'Telemetry' table contains 5 million records. Users frequently filter system views based on the 'Device Category' column. Currently, 'Device Category' is a Lookup column connected to a separate table containing only four records (Mobile, Desktop, Tablet, Server). Users report severe performance degradation when sorting or filtering these views. What is the most performant architectural change to resolve this issue?
check_circle
Correct AnswerC
Lookup columns require the underlying SQL database to perform JOIN operations to retrieve the related data. When querying views with millions of rows, JOINs on Lookups can cause significant performance degradation. Because the category list is extremely small and static (only four options), converting this architecture to a 'Choice' column is highly recommended. Choice columns store simple integer values directly on the record, eliminating the need for complex JOINs and vastly accelerating view filtering and sorting on massive datasets.
Question 113
A Model-driven app user tries to ask the Copilot sidecar to summarize a custom 'Project' record. However, Copilot responds that it cannot find the information or cannot answer the question, even though Dataverse Search is enabled at the environment level. What is the most likely administrative configuration missing?
check_circle
Correct AnswerA
For the built-in Copilot in model-driven apps to reason over data, summarize records, and answer natural language questions, it relies entirely on the Dataverse Search infrastructure. It is not enough to just enable Dataverse Search for the environment; the specific tables (like 'Project') AND the specific columns you want Copilot to understand must be explicitly added to the Dataverse Search index (typically managed by configuring the table's Quick Find view and the environment's search index settings). If it's not indexed, Copilot cannot 'see' it.
Question 114
A healthcare organization wants to automate the intake of patient referral forms. These forms are submitted as PDFs and contain a complex mix of handwritten notes, standard checkboxes, and typed text. Once extracted, the structured data must be automatically saved into a Dataverse 'Referral' table without human intervention. Which combination of components provides the architecturally correct solution for this? (Select two)
Select all that apply
check_circle
Correct AnswersA, C
To automate document intake autonomously (without human intervention), a Power Automate cloud flow is required to orchestrate the process (triggered automatically when the PDF arrives via email or SharePoint). To extract structured key-value pairs from complex PDFs that contain a mix of typed text, checkboxes, and handwriting, the AI Builder 'Document Processing' model (formerly Form Processing) must be trained and called within the flow. The flow then maps the model's outputs to a Dataverse 'Add a row' action.
Question 115
A utility company requires a mobile application for field workers. The workers frequently operate in remote areas without reliable internet access. The application must allow them to conduct equipment inspections, take photos of machinery, and utilize an AI model to identify rusting components directly from the images. Which combination of Power Platform components is most appropriate?
check_circle
Correct AnswerD
Canvas Apps are the ideal choice for building highly customized mobile experiences that require native device hardware integration (like the camera) and, critically, support robust offline capabilities. AI Builder Object Detection models are specifically designed to analyze images and identify/count specific items (like rusting components) within them. Model-driven apps and Power Pages are not optimized for this level of customized, offline, camera-centric mobile interaction, and Text Recognition (OCR) extracts text, not physical objects/rust.
Question 116
You are building a Canvas app with 25 distinct screens. Every screen requires a standard corporate header containing a logo, the logged-in user's name, and a dynamic title that reflects the current screen's name. To minimize maintenance and promote reusability, you decide to build a Canvas Component for this header. How should you pass the dynamic screen name into the Component?
check_circle
Correct AnswerC
Canvas Components use custom properties to communicate with the hosting screen. An 'Input' property allows the host screen to pass data into the component. By defining an Input property called 'ScreenTitle', you can map that property to the Text value of a label within the component. When you place the component on Screen 1, you set the property to 'Home'; on Screen 2, you set it to 'Details'. This ensures the component remains isolated, agnostic, and highly reusable.
Question 117
A retail company wants a mobile app for store managers to audit store displays. Managers must take a photo of the shelves using their mobile device. The system must immediately analyze the photo to identify and count if specific competitor brand products are visible. If competitor products are found, the system must save the data to Dataverse and alert the regional director. Because some stores have zero cell reception or Wi-Fi, offline capability is strictly required for capturing the photo and logging the data. Which implementation approach is required?
check_circle
Correct AnswerC
Canvas Apps provide the robust, customizable offline capabilities required for mobile camera integration in disconnected environments. To detect, locate, and count specific physical items (competitor products) within an image, the AI Builder Object Detection model is the exact AI tool required (Image Classification only tags the whole image, it doesn't count specific objects). Dataverse natively supports Canvas offline caching. A Power Automate flow handles the asynchronous alert (emailing the director) once the app reconnects to the internet and syncs the data to Dataverse.
Question 118
A global enterprise uses Business Units to separate data between the North America, EMEA, and APAC divisions. A strict RBAC security model prevents salespeople in one division from seeing 'Opportunity' records in another. However, the newly appointed Global VP of Sales needs to view every Opportunity owned by any salesperson who reports to them directly or indirectly, regardless of which Business Unit the salesperson resides in. How do you implement this without breaking the Business Unit segregation?
check_circle
Correct AnswerC
Hierarchy Security (specifically the Manager Hierarchy model) operates alongside standard Business Unit Role-Based Access Control (RBAC). When enabled, a manager automatically inherits read access to all records owned by their direct reports, and the records owned by their reports' reports, cascading down the management chain. Crucially, this access traverses Business Unit boundaries, allowing the Global VP to see their entire reporting line's data without granting them blanket 'Organization' level access (which would expose data outside their reporting structure).
Question 119
You are writing a complex `Patch` function in a Canvas app to update multiple Dataverse tables simultaneously. Occasionally, the operation fails because of server-side Dataverse synchronous plugins. You need the app to gracefully catch the failure, prevent a generic error screen, notify the user with the exact error message returned by Dataverse, and update a local variable `varIsProcessing` to false. Which Power Fx implementation achieves this?
check_circle
Correct AnswerB
In Power Fx, formula-level error management is handled using the `IfError` function. When an operation like `Patch` fails inside `IfError`, the function intercepts the failure and executes the fallback arguments. To retrieve the exact server-side error message generated by Dataverse, you must use the `Errors(DataSource)` function, which returns a table of recent errors for that data source. Option 1 correctly intercepts the error, notifies the user with the specific Dataverse message, and executes the state update variable. Try/Catch (Option 2) is C# syntax, not Power Fx.
Question 120
You are designing a Dataverse security model for an intelligent HR application. You must ensure that the 'Salary' column on the 'Employee' table is strictly visible ONLY to users possessing the 'HR Manager' security role. All other employees must be able to view the rest of the employee profile details, but the Salary field must be masked or hidden across the entire platform. How do you enforce this?
check_circle
Correct AnswerC
Column-level security (Field-level security) in Dataverse allows you to restrict read, create, or update access to specific high-business-impact columns (like 'Salary') while keeping the rest of the record accessible. Once enabled on the column, you create a Column Security Profile, grant it read/write permissions for that column, and assign the profile to the 'HR Manager' users. Business Rules only hide fields on the UI (forms) and can be easily bypassed via API or Advanced Find, making them insecure for sensitive data.
