Free Microsoft Power Platform Fundamentals (PL-900) Sample Questions

Correct answer: A — Desktop flow (robotic process automation)

• A. Desktop flows provide robotic process automation that can interact with legacy desktop applications lacking APIs, such as clicking through screens and copying data, which fits this nightly task.
• B. An automated cloud flow triggered by an email would only start the process on a new message; it doesn't provide the UI-level interaction with a legacy desktop app that this task needs.
• C. An instant cloud flow needs someone to manually press a button to start it, which doesn't match the need for an unattended nightly automation.
• D. A scheduled cloud flow can run on a timer, but on its own it cannot interact with an older desktop application's screens the way a desktop flow can.

Correct answer: A — It provides a secure, structured data platform apps, flows, and agents can share

• A. Dataverse gives Power Apps, Power Automate, and Copilot Studio a common, secure, structured store for business data, so different tools can work from the same tables.
• B. Dataverse is a data platform, not a workflow engine; Power Automate remains the tool used to design and run automated flows.
• C. Publishing content to external visitors is the role of Power Pages; Dataverse itself does not render public websites.
• D. Dataverse is a shared data platform usable by Power Apps and Power Automate as well, not a chat surface limited to Copilot Studio.

Correct answer: A — Microsoft Dataverse

• A. Dataverse is designed as the shared data platform so a canvas app, a cloud flow, and a Copilot Studio agent can all read and write the same records without duplicating data.
• B. A custom connector lets a flow or app call an external API, but it isn't itself a place to store shared records.
• C. A desktop flow automates actions on a desktop computer; it doesn't provide a shared data store for multiple tools to read and write.
• D. A Power Pages site is a public-facing website; it consumes data rather than serving as the underlying shared data store for internal apps and flows.

Correct answer: A — Premium connector

• A. Premium connectors require an additional license or per-user/per-flow premium plan beyond what's included with standard Power Apps or Power Automate licensing, which matches a connector needing extra licensing to reach a third-party CRM.
• B. Standard connectors are included in base Power Apps and Power Automate licenses and don't require an additional premium plan.
• C. A custom connector is one a maker builds themselves against an API that doesn't have a prebuilt connector; the scenario doesn't describe building one from scratch.
• D. "Virtual connector" is not a standard Power Platform connector category; it doesn't describe licensing tiers.

Correct answer: A — Lookup

• A. A Lookup column establishes a many-to-one relationship by referencing a row in a related table, such as an order referencing its customer.
• B. Choice provides a fixed set of selectable values but does not reference a row in another table.
• C. Currency stores monetary values along with an associated exchange rate, not a reference to another table.
• D. Yes/No stores a simple two-option boolean value and has no relationship function.

Correct answer: A — On the Contact table, referencing the related Account row

• A. In a 1:N relationship, the lookup column sits on the 'many' side, so each Contact row stores a reference to its related Account.
• B. The Account, on the 'one' side, does not hold a single lookup column pointing to many contacts; it instead has a related list of contacts.
• C. A hidden intersect table is used for many-to-many relationships, not for a one-to-many relationship like this one.
• D. The lookup column exists on only one side of a 1:N relationship, not duplicated on both tables.

Correct answer: A — To define which columns appear and how rows are filtered and sorted when a table's data is listed

• A. A view specifies the columns, filters, and sort order used when a list of table rows is displayed to users, similar to a saved query.
• B. User access to an environment is controlled through security roles and licensing, not through views.
• C. Relationships between tables are defined through relationship configuration, not through views.
• D. Active currencies are configured in organization settings, unrelated to how views control row display.

Correct answers: A, B — Propose new tables and columns from a plain-language description; Suggest relevant columns for a table based on its purpose

• A. Copilot can interpret a maker's description of the data they want to track and propose a matching table structure.
• B. Based on the described purpose, Copilot can suggest relevant columns to include on the generated table.
• C. This AI experience works within the Power Apps maker portal rather than generating raw SQL scripts for a database administrator.
• D. Makers still review and can adjust the proposed structure before it is saved; the process is not designed to skip maker review.

Correct answer: A — To share record ownership and security roles across a group of users

• A. Teams let a group of users share ownership of records and inherit security roles collectively, simplifying access management for a group.
• B. Restricting environment regions relates to environment location settings, not teams.
• C. Classifying connectors as business or non-business is a data loss prevention concept, unrelated to teams.
• D. Scheduling backups is a platform-managed environment operation and is not a function performed by teams.

Correct answer: B — Production environment

• A. The default environment can technically host pipelines, but it is not recommended for most customers as the pipelines host.
• B. A production environment is the recommended choice for the pipelines host because it offers stability and full administrative control.
• C. Trial environments are temporary and expire after 30 days, making them unsuitable as a long-lived pipelines host.
• D. Developer environments are personal, license-tied spaces intended for individual development, not for hosting an organization's pipeline.

Correct answers: A, B, C — The updated policy configuration cascades down to every environment in the tenant; Resources that violate the new policy are placed into a suspended or quarantine state; Connections using a now-blocked connector are set to a disabled state

• A. After a policy is saved, its configuration is cascaded down to each environment in the tenant so resources can be evaluated against it.
• B. Apps, flows, or chatbots found to violate the updated policy are placed into a suspended or quarantine state so they cannot operate.
• C. If a whole connector is blocked, the system scans and disables the connections that rely on it, causing dependent flows to fail.
• D. A DLP violation suspends or quarantines the resource; it does not delete the flow outright.
• E. Enforcement has latency that depends on tenant size, taking up to 24 hours in extreme cases, so it is not instantaneous.

Correct answer: B — False

• A. The default environment is a permanent part of the tenant and cannot be deleted, so this attempt would not succeed.
• B. This is correct: the default environment cannot be deleted, though it can be renamed and its settings adjusted.

Correct answer: A — Canvas app

• A. Canvas apps give makers full control over screen layout, which matches the need for a completely custom, hand-designed interface.
• B. Model-driven apps generate their layout from Dataverse components rather than letting the maker design a fully custom screen.
• C. Copilot Studio agents are conversational assistants, not data entry apps with custom screens for scanning items.
• D. A Power BI dashboard visualizes data but does not provide an interactive data-entry interface for warehouse staff.

Correct answer: B — Professional developers who build apps in a code-first IDE

• A. Citizen developers without coding experience are better served by canvas or model-driven apps rather than a code-first approach.
• B. Code apps let professional developers build apps in code-first IDEs like Visual Studio Code while still running on the managed platform.
• C. Code apps are built and maintained by developers, not simply consumed as static reports by end users.
• D. Business analysts using only drag-and-drop tools are the audience for canvas or model-driven apps, not code apps.

Correct answer: B — Power Apps Premium license

• A. End users cannot run code apps without any license; a specific license tier is required.
• B. End users that run code apps need a Power Apps Premium license.
• C. A time-limited free trial license is not the ongoing licensing requirement for running published code apps.
• D. A Power BI Pro license covers Power BI content and does not grant the right to run code apps.

Correct answers: A, C, D — Connecting to hundreds of data sources including Dataverse, SharePoint, and Excel; Creating responsive designs that work across browsers and mobile devices; Using AI-powered Copilot to build the app through natural language conversation

• A. Canvas apps can connect to hundreds of data sources, giving makers broad flexibility in choosing where data comes from.
• B. Building from a fixed data model with no layout control describes model-driven apps, not canvas apps.
• C. Canvas apps are designed to create responsive experiences that work across browsers and mobile devices.
• D. Canvas apps support Copilot-assisted, conversational app creation in addition to manual design.
• E. Canvas apps are aimed at low-code makers and do not require writing React code for every screen.
• F. Canvas apps run as standalone business applications and are not limited to running inside Power BI.

Correct answer: A — Start with data, then select Create new data

• A. Selecting Start with data and then Create new data opens the Create new tables workspace, where the Copilot panel lets a maker describe the data they want and generate Dataverse tables for their app.
• B. Starting from a blank canvas begins a manually designed app with no screens or data, so it does not launch the Copilot table-generation workflow.
• C. Importing a solution brings in existing components rather than starting the natural-language, Copilot-driven table creation experience.
• D. Creating a flow opens Power Automate's flow designer, which builds automation, not the Dataverse tables used to start a canvas app with Copilot.

Correct answer: B — False

• A. This describes canvas apps, where makers have pixel-level control, not the data-first design of model-driven apps.
• B. Model-driven apps generate their forms, views, and navigation from the data model, so makers configure and adjust rather than laying out every pixel.

Correct answers: A, B, C — Canvas and model-driven apps aren't supported in the new authoring experience; Only one app per plan is supported; Coauthoring by multiple makers at the same time isn't supported

• A. The new authoring experience doesn't support opening, recommending, or editing canvas or model-driven apps.
• B. Currently, a plan can only be linked to a single app; multiple apps per plan aren't supported.
• C. Multiple makers editing the same app at the same time, or coauthoring, isn't supported.
• D. Editing existing Dataverse tables through chat isn't currently supported; changes need to be made manually in the data editor.
• E. Plans created before or outside of vibe can't be opened in the new experience.

Correct answer: B — False

• A. Coauthoring by multiple makers editing the same app simultaneously is explicitly listed as unsupported.
• B. This is correct: simultaneous multi-maker editing, or coauthoring, isn't supported in Power Apps vibe.

Correct answers: A, B, D — The automation needs to connect multiple online services and apps through connectors; The automation must run in a cloud environment without depending on a specific local machine; The automation should trigger when an email arrives from a specific sender

• A. Cloud flows are built to connect apps and services through connectors, making this a valid reason to prefer a cloud flow.
• B. Cloud flows run in the cloud and do not depend on a specific desktop being online, which is a valid reason to choose them over desktop flows.
• C. Clicking through a legacy desktop application with no API is precisely the robotic process automation use case that desktop flows, not cloud flows, are designed for.
• D. Triggering on the arrival of an email from a specific sender is a classic automated cloud flow scenario, supporting cloud flow use.
• E. Recording mouse movements on a local screen is a UI automation capability specific to desktop flows, not a reason to choose a cloud flow.

Correct answer: B — False

• A. This is incorrect; desktop flows are specifically capable of automating legacy applications like terminal emulators in addition to modern web and desktop apps.
• B. Desktop flows support legacy applications such as terminal emulators alongside modern web and desktop applications, so the claim that they cannot is false.

Correct answer: A — True

• A. Scheduled cloud flows are designed exactly for predictable recurring tasks, letting a single flow definition run repeatedly instead of building a new flow for every occurrence.
• B. This statement correctly describes scheduled cloud flow behavior, so marking it false would be incorrect.

Correct answer: A — An automated cloud flow triggered by new Forms responses that writes to SharePoint, with an approval step for flagged entries

• A. Triggering automatically on new Forms responses eliminates the manual inbox check, writing directly to SharePoint, while an approval step preserves human review for unusual or flagged entries.
• B. Running once a year to archive old forms does not address the daily need to process new enrollment submissions promptly.
• C. A desktop flow clicking through Outlook manually every morning still requires scheduling UI interaction and does not use the more direct Forms trigger available for this scenario.
• D. A business process flow for sales opportunities is a different Power Platform feature focused on guiding sales stages, not on processing enrollment form submissions.

Correct answer: A — Scheduled cloud flow

• A. A scheduled cloud flow is correct because it runs at a specific time or on a recurrence you define, which matches the every-Monday-at-8-AM requirement.
• B. An instant cloud flow requires a person to manually start it, which does not match a flow that must run automatically on a fixed weekly schedule.
• C. An automated cloud flow responds to an event in a connected service rather than to the passage of time on a calendar schedule.
• D. A desktop flow automates tasks on a local machine and is not the flow type used to define a recurring calendar-based schedule.

Correct answer: A — Copilot's flow-creation feature for desktop flows is limited to US environments

• A. This is correct because, at the time of writing, the flow-creation capability of Copilot in Power Automate for desktop is only available in environments located in the United States, so a German environment would not have access.
• B. The account requirement runs the other way: a work or school account is required, and this user already meets that requirement, so the account is not the cause.
• C. The feature has not been discontinued; it remains available, but only within supported regions such as the United States.
• D. Desktop flows still support other Copilot capabilities, such as answering product questions, even where flow creation is not yet available.

Correct answer: A — Record with Copilot

• A. Record with Copilot is correct because it is designed for exactly this scenario: sharing your screen and narrating the process so that Power Automate interprets your voice and actions into an automation.
• B. The classic Recorder captures clicks and keystrokes to repeat them, but it does not use narration and cannot infer logic like conditions from spoken explanation.
• C. A scheduled cloud flow runs on a time-based recurrence and has nothing to do with demonstrating a desktop process on screen.
• D. The security role editor manages user permissions in an environment and is unrelated to building an automation from a screen demonstration.

Correct answer: B — False

• A. This is not true; personal Microsoft accounts are explicitly excluded from using Copilot in Power Automate for desktop.
• B. This is false, because Copilot in Power Automate for desktop requires a work or school account, not a personal account like a live.com address.

Correct answers: A, B — Question node; Condition node

• A. A Question node asks the customer a question and captures their response, which is the first step needed to branch on an answer.
• B. A Condition node branches the conversation based on a condition, such as the value captured from the prior question, making it the mechanism for the branching itself.
• C. A Message node only sends a message to the customer; it does not capture input or branch the conversation.
• D. Variable management nodes set, parse, or clear values but do not themselves ask a question or branch the conversation path.
• E. A Trigger node starts a topic based on recognized intent; it does not ask a follow-up question or branch based on an answer.

Correct answer: A — Defining a REST API tool that selects specific endpoints and methods to call

• A. Copilot Studio supports defining a REST API tool directly, letting a maker select endpoints and methods to call an API without first building a custom connector.
• B. Uploading documentation as a knowledge source lets the agent describe the API in text but does not let it actually call the API to retrieve live order status.
• C. Makers cannot create new system topics, and hardcoding status values would not reflect real, current order data.
• D. Publishing to another channel affects where users can reach the agent but does not connect the agent to the internal API.

Correct answer: B — False

• A. This is incorrect because customers only see the latest content after the maker publishes again; unpublished changes are not visible to users on connected channels.
• B. This is correct: the maker must publish again after any change for that change to reach users on the connected channels, since publishing applies updates across all channels together.

Correct answer: A — The flow is missing the When an agent calls the flow trigger, the Respond to the agent action, or has not been published

• A. A flow only appears as an available tool if it has the correct trigger, a real-time Respond to the agent action, and has been published; missing any of these keeps it off the list.
• B. Reaching the 128-tool orchestrator maximum is possible but is a far less common cause than a flow simply not meeting the basic trigger, response, and publish requirements.
• C. Building the flow with natural language versus the visual designer does not affect whether it can later be added as a tool, as long as the structural requirements are met.
• D. Channel configuration for publishing the agent is unrelated to whether an agent flow appears in the tool selection list.

Correct answer: A — Microsoft Defender

• A. Microsoft Defender contributes threat protection and vulnerability management capabilities that Agent 365 relies on to secure agents.
• B. Planner is a task-management app and provides no threat protection or agent security capability.
• C. Viva Insights focuses on workplace productivity analytics, not agent threat protection.
• D. Bookings schedules appointments and has no connection to agent security or vulnerability management.

Correct answer: A — 30 minutes

• A. A conversation times out after 30 minutes of inactivity for channels other than Telephony, which affects how sessions are counted in analytics.
• B. The 3-minute timeout applies specifically to the Telephony channel after an End Conversation event, not to general channels.
• C. A full day is far longer than the actual inactivity timeout used for conversation sessions.
• D. A week is far longer than the actual inactivity timeout and does not match documented conversation session behavior.

Correct answer: A — The customer comments summary, which prioritizes and groups related feedback by volume and significance

• A. The customer comments summary is built to surface customer sentiment by grouping related feedback and prioritizing insights by volume and significance, exactly this scenario.
• B. Topic outcomes reports categorical resolution results, not free-text sentiment about wait times.
• C. Active users tracks engagement counts, not the content or sentiment of customer comments.
• D. Downloading transcripts gives raw logs but does not automatically summarize or prioritize sentiment across many conversations.