PL-900 mock exam: 25 free practice questions

Q1 / 25

A generated desktop flow contains a placeholder where Copilot identified browser work, and that placeholder shows a design-time error. The maker cannot run the recorder on this machine today but must hand the flow over without errors. What should the maker do?

1. ADelete the placeholder, add the browser actions manually, and hand over the completed flow
2. BLeave the placeholder in place, because the design-time error clears once the flow is saved
3. CHand the flow over as it is, because a placeholder is simply skipped when the flow runs
4. DDelete the whole generated flow and describe the scenario again without the browser step

Q2 / 25

A generated desktop flow contains a placeholder where Copilot identified browser work, and that placeholder shows a design-time error. The maker cannot run the recorder on this machine today but must hand the flow over without errors. What should the maker do?

1. ADelete the placeholder, add the browser actions manually, and hand over the completed flow
2. BLeave the placeholder in place, because the design-time error clears once the flow is saved
3. CHand the flow over as it is, because a placeholder is simply skipped when the flow runs
4. DDelete the whole generated flow and describe the scenario again without the browser step

Q3 / 25

Which set of elements is controlled by a Dataverse view?

1. ARecord tabs, sections, labels, and embedded timelines
2. BColumn data types, validation rules, Power Fx formulas, and requirement levels
3. CFlow triggers, actions, approvals, and time schedules
4. DDisplayed records, columns, filter conditions, and sort order

Q4 / 25

An existing flow ends by sending an Outlook email once an expense is approved, and the team now wants a Teams message instead. The trigger and the approval must stay exactly as they are. How is that change made with Copilot?

1. ADescribe the whole process again in the home page field and keep the newer proposal
2. BOpen the flow in the designer and ask its Copilot chat panel to replace that action
3. CAsk the Copilot chat panel to turn the approval step into a Teams connector trigger
4. DEdit the output of the email action so that it posts into the Teams channel instead

Q5 / 25

A compliance answer may be worded freely but must always contain the phrase manager approval required, and the team has to keep proof of a passing run for an audit eight months away. What fits both needs?

1. AEvaluate with tool use, and export the results to CSV while they are still available
2. BEvaluate with exact match, and export the results to CSV while they are still available
3. CEvaluate with keyword match, and export the results to CSV while they are still available
4. DEvaluate with keyword match, and leave the results in Copilot Studio until the audit asks

Q6 / 25

Which Power Platform admin center area focuses on solution health and near-real-time failures?

1. ADeployment, which reports application session trends
2. BSecurity, which configures solution deployment stages and pipeline approvals
3. CMonitor, which surfaces health and performance problems
4. DActions, which stores long-term application telemetry

Q7 / 25

Which Power Platform admin center area focuses on solution health and near-real-time failures?

1. ADeployment, which reports application session trends
2. BSecurity, which configures solution deployment stages and pipeline approvals
3. CMonitor, which surfaces health and performance problems
4. DActions, which stores long-term application telemetry

Q8 / 25

What does Copilot generate for configured tables in an initial model-driven app?

1. AForms, views, and navigation
2. BA published app with no review stage
3. CBrowse, detail, and edit canvas screens
4. DA blank interface without Dataverse tables

Q9 / 25

Which four responsibilities summarize Microsoft Agent 365 capabilities?

1. AControl, observe, secure, and govern agents
2. BImport, score, retain, and export test cases
3. CRetrieve, augment, generate, and cite answers
4. DPrompt, branch, redirect, and publish topics

Q10 / 25

A pipeline stage must let makers request a release into production without giving them access to it, hold the release until a named reviewer accepts, and check the solution before anything is imported. Which configuration covers all three?

1. AMaker-identity deployment with an approval gate, so that only an approved maker reaches production
2. BDelegated deployment, an approval gate, and the pipeline's own preflight checks
3. CDelegated deployment and an approval gate, with Solution Checker run by the maker before requesting
4. DDelegated deployment and preflight checks, with the reviewer notified once the deployment has finished

Q11 / 25

A pipeline stage must let makers request a release into production without giving them access to it, hold the release until a named reviewer accepts, and check the solution before anything is imported. Which configuration covers all three?

1. AMaker-identity deployment with an approval gate, so that only an approved maker reaches production
2. BDelegated deployment, an approval gate, and the pipeline's own preflight checks
3. CDelegated deployment and an approval gate, with Solution Checker run by the maker before requesting
4. DDelegated deployment and preflight checks, with the reviewer notified once the deployment has finished

Q12 / 25

Which refinement is supported before Copilot generates a model-driven app?

1. APositioning every control on a blank canvas screen
2. BPublishing the app before reviewing its generated tables
3. CRemoving the data model and retaining sample records
4. DEstablishing relationships between generated Dataverse tables

Q13 / 25

A commission calculation must apply the same rate table in the same order every time, be auditable, and return one amount to the agent. A second task lets technicians ask about their ServiceNow tickets in their own words, with the agent deciding which operations to call. Which pair fits?

1. AGenerative orchestration for both, letting the model decide the order in either case
2. BAn agent flow for both, with its steps rewritten for each kind of ticket question
3. CAn MCP server for the commission, and an agent flow for the ticket questions
4. DAn agent flow for the commission, and an MCP server connection for the ticket questions

Q14 / 25

An existing flow ends by sending an Outlook email once an expense is approved, and the team now wants a Teams message instead. The trigger and the approval must stay exactly as they are. How is that change made with Copilot?

1. ADescribe the whole process again in the home page field and keep the newer proposal
2. BOpen the flow in the designer and ask its Copilot chat panel to replace that action
3. CAsk the Copilot chat panel to turn the approval step into a Teams connector trigger
4. DEdit the output of the email action so that it posts into the Teams channel instead

Q15 / 25

A finance process must start when an invoice email arrives, ignore the newsletters that reach the same mailbox, and then enter the data into an application that has no connector. Which design fits?

1. AA scheduled hourly cloud flow that opens the mailbox with a desktop flow and reads what arrived
2. BAn automated cloud flow on the mail event that enters the data with a connector action instead
3. CAn instant cloud flow that a finance clerk starts for each invoice, calling the same desktop flow
4. DAn automated cloud flow on the mail event, with a condition on the message, calling a desktop flow

Q16 / 25

Which set can appear in a solution generated by Plans?

1. AOne canvas screen and its data source, with the roles and requirements kept separate
2. BDataverse tables, apps, cloud flows, sites, reports, and agents
3. CReact source files, npm packages, and local debug logs, generated from the plan
4. DDataverse tables and reports, with apps, flows, and agents proposed separately afterwards

Q17 / 25

Which set can appear in a solution generated by Plans?

1. AOne canvas screen and its data source, with the roles and requirements kept separate
2. BDataverse tables, apps, cloud flows, sites, reports, and agents
3. CReact source files, npm packages, and local debug logs, generated from the plan
4. DDataverse tables and reports, with apps, flows, and agents proposed separately afterwards

Q18 / 25

How does Microsoft Agent 365 relate to existing organizational controls?

1. AIt removes existing controls and gives each agent an isolated governance model
2. BIt extends existing management, security, and governance processes to agents
3. CIt limits existing controls to conversation topics in one published channel
4. DIt converts existing controls into ungoverned MCP operations at runtime

Q19 / 25

A finance process must start when an invoice email arrives, ignore the newsletters that reach the same mailbox, and then enter the data into an application that has no connector. Which design fits?

1. AA scheduled hourly cloud flow that opens the mailbox with a desktop flow and reads what arrived
2. BAn automated cloud flow on the mail event that enters the data with a connector action instead
3. CAn instant cloud flow that a finance clerk starts for each invoice, calling the same desktop flow
4. DAn automated cloud flow on the mail event, with a condition on the message, calling a desktop flow

Q20 / 25

Where does a maker begin conversational creation of a Dataverse table?

1. ASelect Flows and then create a scheduled cloud flow
2. BSelect Views and then create an Active records query with columns and sorting
3. CSelect Forms and then add a Quick create layout
4. DSelect Tables and then Start with Copilot in the maker portal

Q21 / 25

A customer-facing portal needs complete control over its design and has to integrate with a proprietary legacy system. The developers want to keep their framework and local debugging, and the organization wants the portal to use platform identity and connectors. Which option meets all of it?

1. AA code app built locally with the team's framework and run on Power Platform
2. BA canvas app with custom controls, which offers the same level of design control
3. CA model-driven app, whose generated interface can be restyled to match the portal design
4. DA custom web app hosted outside Power Platform, since identity and connectors are not open to code

Q22 / 25

A customer-facing portal needs complete control over its design and has to integrate with a proprietary legacy system. The developers want to keep their framework and local debugging, and the organization wants the portal to use platform identity and connectors. Which option meets all of it?

1. AA code app built locally with the team's framework and run on Power Platform
2. BA canvas app with custom controls, which offers the same level of design control
3. CA model-driven app, whose generated interface can be restyled to match the portal design
4. DA custom web app hosted outside Power Platform, since identity and connectors are not open to code

Q23 / 25

Which analytics view helps identify conversation areas with frequent triggers and high escalation or abandonment?

1. ACustomer Satisfaction
2. BTopic analytics
3. CActive Users
4. DAgent Inventory

Q24 / 25

What does Copilot generate for configured tables in an initial model-driven app?

1. AForms, views, and navigation
2. BA published app with no review stage
3. CBrowse, detail, and edit canvas screens
4. DA blank interface without Dataverse tables

Q25 / 25

What distinguishes a delegated pipeline deployment?

1. AIt runs under every maker's union of environment security roles
2. BIt skips solution validation for deployments to sensitive environments
3. CIt runs under a service principal or pipeline stage owner's identity
4. DIt replaces the pipeline host with an analytics data lake