Q01 - Question
A semantic model supports sales, finance, and service report areas. You want Copilot to use only the sales fields when it answers data questions. The finance and service reports must continue to use their fields in the same semantic model. You use Prep data for AI in Power BI Desktop. Which control should you use?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. An AI data schema that limits Copilot to the sales fields.
- B. An AI instruction that tells Copilot to ignore finance and service fields.
- C. A verified answer for each sales question.
- D. Deleting the finance and service fields from the semantic model.
A is correct.
Explanation: An AI data schema defines the focused fields that Copilot uses for data questions. Use it when a semantic model supports multiple report areas but Copilot should use only a subset of fields. It limits Copilot to the sales fields without removing fields from the semantic model.
B is incorrect: An AI instruction adds business rules that the semantic model cannot express, such as which city field to use by default. It is not the control for limiting which fields Copilot considers.
C is incorrect: A verified answer returns a human-approved visual and filters when a prompt matches a trigger phrase. It fits a common question that needs the same answer every time. It does not limit the fields that Copilot uses for other questions.
D is incorrect: Deleting the fields would remove them from the semantic model, and the finance and service report areas need them. An AI data schema limits Copilot without removing fields.
Q02 - Question
You test a semantic model with priority questions before you mark it Approved for Copilot. In one test, Copilot includes a technical location field in an answer instead of Customer City. In another test, Copilot does not recognize a business phrase that users commonly use for the Total Sales measure. Select the two actions that address these two findings.
Domain: Implement and manage semantic models (25–30%) Type: Multi-select
- A. Create a verified answer to replace the technical field problem.
- B. Remove the technical field from the AI data schema.
- C. Approve the semantic model because the other priority questions returned the expected numbers.
- D. Add the business phrase as a synonym for Total Sales.
B and D are correct.
Explanation: A technical field in an answer is fixed by removing it from the AI data schema. An unknown business phrase is fixed by adding a synonym. After each fix, close and reopen the Copilot pane and ask the same question again to confirm that the fix works without breaking other answers.
A is incorrect: A verified answer is for a stable question that needs a set answer. It does not remove a technical field from the fields that Copilot considers.
C is incorrect: Approve the semantic model only after priority questions consistently use the trusted definitions. Approval without fixing and retesting the failed prompts means every Copilot experience that uses the model inherits untested behavior.
Q03 - Question
You created a contains relationship type with Hospital as the origin entity type and Department as the target entity type. You must make it queryable so the ontology shows which hospital contains which departments. The departments table has the columns DepartmentId and HospitalId. The hospitals table has only HospitalId and hospital attributes. How should you configure the relationship?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Use the hospitals table as the source and match HospitalId for the Hospital entity type.
- B. Leave the relationship without source data. The relationship type definition is enough to make connections queryable.
- C. Use the departments table as the source and match only DepartmentId, because HospitalId is already defined on Hospital.
- D. Use the departments table as the source, match HospitalId to Hospital, and match DepartmentId to Department.
D is correct.
Explanation: A relationship binding needs a source table that identifies both entity types. You select a matched column for each entity type, which is the column that matches that entity type's key. The departments table has HospitalId for the origin and DepartmentId for the target. Each row with values in both columns creates one relationship instance.
A is incorrect: The hospitals table has no column that identifies a department, so it cannot connect specific hospitals to specific departments.
B is incorrect: A relationship type without a binding is only a definition and cannot be queried.
C is incorrect: You select a matched column for each entity type. Without HospitalId, rows do not link a department to a specific hospital.
Q04 - Question
Source data for a semantic model uses a snowflake schema: Product, Subcategory, and Category. Both the Sales fact table and the Inventory fact table use the same Category table. The team is concerned that denormalizing Category into separate dimension copies would create inconsistent data. What should the team do?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Preserve the snowflake structure and use single-direction filtering from Category toward the fact tables.
- B. Flatten the snowflake into one Product dimension and keep a separate copy of the category columns for each fact table.
- C. Preserve the snowflake structure and use bi-directional filtering on every relationship in the chain.
- D. Remove the Category table from the model so that each fact table has fewer relationships.
A is correct.
Explanation: Preserve a snowflake structure when multiple fact tables share a subdimension table and denormalization would create inconsistent copies. Configure each relationship in the chain with single-direction filtering from the outermost table toward the fact table. A filter on Category then flows through Subcategory and Product into the fact table.
B is incorrect: Flattening is usually the better choice, but here it would create separate copies of the shared category data for each fact table. That is the inconsistency the team wants to avoid.
C is incorrect: Bi-directional filtering should be used sparingly. It can degrade query performance and create unexpected filter behavior in reports. The standard pattern for a preserved snowflake is single-direction filtering.
D is incorrect: Removing Category would remove the attribute that both fact tables need for slicing. It does not address the shared-dimension requirement.
Q05 - Question
Your team builds a new semantic model that uses Direct Lake storage mode on large Delta tables in OneLake. No performance tests have been run yet. A team member proposes to create aggregation tables for every large table before the model is released. What should you recommend?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Enable automatic aggregations training in the semantic model settings, because automatic aggregations are required for all large Delta tables.
- B. Change the model to DirectQuery storage mode so that you can create user-defined aggregations.
- C. Test the model performance without aggregations first, and add aggregations only if the measurements show a clear need.
- D. Create aggregation tables for all large Delta tables now, so that every query reads summarized data.
C is correct.
Explanation: The evidence supports C because Direct Lake is already fast for many scenarios as it avoids the overhead of data import and external query round-trips. The recommended practice is to start by testing performance without aggregations and add them only when measurement shows a clear need.
A is incorrect: Automatic aggregations require DirectQuery storage mode and are not required for all large Delta tables in Direct Lake mode.
B is incorrect: Changing the storage mode to DirectQuery would remove the performance benefits of Direct Lake and is not recommended before testing the baseline performance.
D is incorrect: Creating aggregation tables for all large tables preemptively adds unnecessary complexity and overhead without verifying if a performance bottleneck actually exists.
Q06 - Question
An analyst has created an ontology item. The analyst now needs to start at maintenance order M001 and traverse relationships to find the dependent task and project. The analyst also wants to write GQL queries for more complex analysis, without data duplication or ETL. Which Fabric IQ item should the analyst use?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Graph in Microsoft Fabric
- B. An operations agent
- C. Planning
- D. A data agent
A is correct.
Explanation: The evidence states that when you create an ontology item, a managed graph is automatically created from your ontology's entity types and relationships. You query it using GQL, and Graph in Microsoft Fabric operates directly on OneLake without requiring data duplication or ETL.
B is incorrect: An operations agent monitors real-time data and recommends or triggers actions based on rules, rather than providing a GQL query interface.
C is incorrect: Planning enables collaborative planning, forecasting, and reporting, rather than traversing relationships with GQL.
D is incorrect: A data agent answers natural language questions and generates SQL, DAX, or KQL, not GQL queries that you write.
Q07 - Question
You built an ontology manually. Equipment attributes such as EquipmentId, PatientId, and EquipmentType are in a lakehouse table. VitalSignsReadings, with EquipmentId, Timestamp, HeartRate, and OxygenSaturation, is in an eventhouse. You need the VitalSignEquipment entity type to expose both the attributes and the readings. Which TWO actions should you take?
Domain: Implement and manage semantic models (25–30%) Type: Multi-select
- A. Create the static data binding for VitalSignEquipment from the lakehouse table first.
- B. Create the time series binding first so the readings create the equipment entities.
- C. Add a time series binding to the VitalSignsReadings table, select Timestamp as the timestamp column, and use EquipmentId as the static linking property.
- D. Add a second static data binding to VitalSignEquipment that points to the eventhouse table.
A and C are correct.
Explanation: The static binding creates the equipment entities, and the time series binding attaches readings to them. The system uses EquipmentId from the static binding to match each reading to the correct equipment entity. The time series binding requires a timestamp column. In it, EquipmentId is the static linking property, and HeartRate and OxygenSaturation are the Timeseries properties.
B is incorrect: Time series data needs existing entities to attach to. Without the static binding, the system has no entity to attach readings to.
D is incorrect: Each entity type supports exactly one static data binding. Eventhouse readings are added through a time series binding.
Q08 - Question
Lamna Healthcare has an existing Power BI semantic model in Import mode that contains hospitals, departments, rooms, and patients. The team wants to start a Fabric IQ ontology from this model to save design time. What should the team expect after generating the ontology?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Entity types, properties, relationship types, and data bindings are all generated automatically.
- B. Only entity types are generated. The team must create properties and relationship types manually.
- C. Entity types, properties, and relationship type definitions are generated. The team must configure data bindings manually.
- D. Generation is not supported for Import mode. The team must convert the model to Direct Lake mode first.
C is correct.
Explanation: Ontology generation works with Import, Direct Lake, and DirectQuery semantic models. All three modes produce entity types, properties, and relationship type definitions. Only Direct Lake mode also generates data bindings, so with Import mode you configure the data bindings yourself after generation.
A is incorrect: Automatic data bindings are generated only for Direct Lake semantic models, not for Import mode.
B is incorrect: Each table becomes an entity type, each column becomes a property, and each model relationship becomes a relationship type.
D is incorrect: Import mode is supported for ontology generation. You do not need to convert the model to Direct Lake mode first. Direct Lake mode only adds automatic data bindings.
Q09 - Question
A report page loads slowly. In Performance analyzer, you identify the slowest visual, a table with many measures. You copy its DAX query and run it in DAX query view. The query returns results quickly, but the visual is still slow to load on the report page. What should you do next?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Simplify the visual by reducing the number of measures or data points, or by applying filters such as Top N.
- B. Rewrite the measures with variables to remove repeated subexpressions.
- C. Reduce the cardinality of the key columns in the model.
- D. Verify query folding in Power Query for the DirectQuery tables.
A is correct.
Explanation: The evidence supports A because if a DAX query runs quickly in DAX query view but the visual is slow to load, the bottleneck is visual rendering. Simplifying the visual by reducing measures, data points, or applying Top N filters addresses rendering issues.
B is incorrect: Rewriting measures with variables optimizes the DAX query time, but the DAX query is already running quickly in this scenario.
C is incorrect: Reducing cardinality improves model size and DAX query performance, but it does not resolve a visual rendering bottleneck.
D is incorrect: Verifying query folding addresses slow DirectQuery source performance, which would manifest as a slow DAX query, not a slow visual rendering time.
Q10 - Question
A fault is recorded in onboard equipment that is carried by vessel V001 but is not required by any scheduled work. An analyst must find which installation tasks and projects depend on equipment that is affected by maintenance order M001. The analyst must not flag every task that is assigned to the vessel. Which traversal should the analyst use?
Domain: Implement and manage semantic models (25–30%) Type: Single choice
- A. Follow carries from Vessel to Equipment, and then follow assignedTo in reverse to find every task on that vessel.
- B. Start at the Project, follow hasTask to every InstallationTask, and then follow installs to Foundation.
- C. Follow affects from MaintenanceOrder to Equipment, follow requiresEquipment in reverse to the InstallationTask, and then follow hasTask in reverse to the Project.
- D. Follow assignedTo from InstallationTask to Vessel, and then treat all equipment on that vessel as required.
C is correct.
Explanation: The evidence states that to investigate a dependency, you follow the maintenance order through affects to the equipment, traverse requiresEquipment in reverse to find the task, and then hasTask in reverse to find the project.
A is incorrect: A vessel carrying equipment is different from a task requiring it. This path would flag tasks that do not depend on the affected equipment.
B is incorrect: This path starts from the project and does not use the maintenance order or the equipment dependency to find affected tasks.
D is incorrect: Treating all vessel equipment as required ignores the requiresEquipment relationship and can flag tasks incorrectly.