cert
cert-1
cert-2

Pass Microsoft PL-400 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
PL-400 Exam - Verified By Experts
PL-400 Premium Bundle
$14.99

PL-400 Premium Bundle

$74.99
$89.98
  • Premium File 429 Questions & Answers. Last update: Nov 10, 2026
  • Training Course 106 Video Lectures
 
$89.98
$74.99
accept 41 downloads in last 7 days
block-screenshots
Purchase Microsoft PL-400 Exam Training Products Individually
Microsoft PL-400 Premium File
Premium File 429 Q&A
$76.99$69.99
Microsoft PL-400 Video Training Course
Training Course 106 Lectures
$21.99 $19.99
Microsoft PL-400 Practice Test 18

Question 170

A header policy accidentally forwards a bearer token to a third-party hostname because its destination rewrite is too broad. Which TWO complementary steps provide the intended implementation AND the evidence needed to validate it? Select TWO.

  1. Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas.
  2. Use durable checkpoints, conflict detection and explicit deletion reconciliation to keep external systems consistent with Dataverse.
  3. Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references.
  4. Inspect delivery attempt logs, event ID, consumer receipt store, downstream transaction and retry/backoff handling.
  5. An incremental sync catches updates and deletions without rescanning every row or missing changes during paging.

Explanation

Correct Answers

A:

Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas. This choice implements the controlling platform requirement rather than masking the symptom. Policy templates can transform requests but cannot substitute for server-side permission enforcement or secret management. Verify the change using: Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references. This is the implementation half of the required two-step answer.

C:

Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references. These observations separate this failure from adjacent capabilities or permissions. Policy templates can transform requests but cannot substitute for server-side permission enforcement or secret management. The relevant implementation is: Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas. This is the diagnostic/verification half of the required two-step answer.

Incorrect Answers

B:

Use durable checkpoints, conflict detection and explicit deletion reconciliation to keep external systems consistent with Dataverse. This step belongs to Incremental sync conflict and deletion recovery rather than the complementary pair here. The required action is: Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas. And its evidence is: Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references.

D:

Inspect delivery attempt logs, event ID, consumer receipt store, downstream transaction and retry/backoff handling. This step belongs to Dataverse event endpoint deduplication and retry rather than the complementary pair here. The required action is: Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas. And its evidence is: Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references.

E:

An incremental sync catches updates and deletions without rescanning every row or missing changes during paging. This step belongs to Dataverse change tracking incremental synchronization rather than the complementary pair here. The required action is: Configure OAuth or supported API key authentication and policy templates for runtime headers/URLs without placing secrets in app formulas. And its evidence is: Inspect authorization URL, scopes, redirect URI, token audience, policy template rules and deployment connection references.

Question 171

A build succeeds but its flow connections and custom APIs fail in test, while the team wants to approve production automatically. Which TWO complementary steps provide the intended implementation AND the evidence needed to validate it? Select TWO.

  1. The rule works for supported users without maintaining custom code and the exception cases remain correct.
  2. Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts.
  3. Check runtime customization needs, agent tools, API authentication, connector or MCP interface, usage monitoring and data access policy.
  4. Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests.
  5. Keep related Dataverse tables, flows, connection references and plug-in registrations in a versioned solution and resolve dependencies before import.

Explanation

Correct Answers

B:

Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts. This choice implements the controlling platform requirement rather than masking the symptom. CI/CD automates deploys but does not supply missing permissions, connection references or application acceptance tests. Verify the change using: Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests. This is the implementation half of the required two-step answer.

D:

Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests. These observations separate this failure from adjacent capabilities or permissions. CI/CD automates deploys but does not supply missing permissions, connection references or application acceptance tests. The relevant implementation is: Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts. This is the diagnostic/verification half of the required two-step answer.

Incorrect Answers

A:

The rule works for supported users without maintaining custom code and the exception cases remain correct. This step belongs to Out-of-box table/form choice versus server extension rather than the complementary pair here. The required action is: Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts. And its evidence is: Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests.

C:

Check runtime customization needs, agent tools, API authentication, connector or MCP interface, usage monitoring and data access policy. This step belongs to Microsoft Foundry versus Power Platform agent boundary rather than the complementary pair here. The required action is: Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts. And its evidence is: Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests.

E:

Keep related Dataverse tables, flows, connection references and plug-in registrations in a versioned solution and resolve dependencies before import. This step belongs to Solution component dependency management rather than the complementary pair here. The required action is: Use Power Platform Pipelines or supported Build Tools with connection/environment settings, solution checker and approvals to release repeatable artifacts. And its evidence is: Inspect PAC CLI export/import steps, environment permissions, build logs, package version, preflight checks and postdeployment smoke tests.

Question 172

A nightly flow stops because it uses a departed maker's personal Dataverse connection. Which TWO complementary steps provide the intended implementation AND the evidence needed to validate it? Select TWO.

  1. Use grouped Scopes, configure run-after branches for failure/timeout and record business state before notifying success.
  2. Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling.
  3. Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors.
  4. Inspect request URI, OData filters, prefer header, continuation link, lookup annotation and calling user privileges.
  5. The agent reads authorized records and cannot modify a customer account through an undeclared or unapproved tool.

Explanation

Correct Answers

B:

Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling. These observations separate this failure from adjacent capabilities or permissions. A service principal with a connector connection still needs Dataverse security roles; retries should not duplicate business effects. The relevant implementation is: Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors. This is the diagnostic/verification half of the required two-step answer.

C:

Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors. This choice implements the controlling platform requirement rather than masking the symptom. A service principal with a connector connection still needs Dataverse security roles; retries should not duplicate business effects. Verify the change using: Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling. This is the implementation half of the required two-step answer.

Incorrect Answers

A:

Use grouped Scopes, configure run-after branches for failure/timeout and record business state before notifying success. This step belongs to Power Automate Scope run-after and failure handling rather than the complementary pair here. The required action is: Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors. And its evidence is: Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling.

D:

Inspect request URI, OData filters, prefer header, continuation link, lookup annotation and calling user privileges. This step belongs to Dataverse OData Web API select filter and pagination rather than the complementary pair here. The required action is: Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors. And its evidence is: Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling.

E:

The agent reads authorized records and cannot modify a customer account through an undeclared or unapproved tool. This step belongs to Microsoft Foundry agent tool access via MCP servers rather than the complementary pair here. The required action is: Use a Dataverse connection authorized by a service principal/application user for unattended flows and configure controlled retries on transient errors. And its evidence is: Inspect connection reference, authentication type, app-user table rights, retry count and 429 handling.

Question 173

Two registrations of the same Create handler run on one record and duplicate external notifications. Which TWO complementary steps provide the intended implementation AND the evidence needed to validate it? Select TWO.

  1. The unattended flow continues after an employee leaves and fails safely on permission or data errors.
  2. Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.
  3. Inspect output parameters, stage, post image configuration, async mode, transaction state and external dependencies.
  4. Define a Dataverse custom API with request/response properties, allowed caller, binding type and registered plug-in implementation.
  5. Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution.

Explanation

Correct Answers

B:

Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer. These observations separate this failure from adjacent capabilities or permissions. Publishing a DLL without a registered step does not automatically attach code to Dataverse messages. The relevant implementation is: Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution. This is the diagnostic/verification half of the required two-step answer.

E:

Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution. This choice implements the controlling platform requirement rather than masking the symptom. Publishing a DLL without a registered step does not automatically attach code to Dataverse messages. Verify the change using: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer. This is the implementation half of the required two-step answer.

Incorrect Answers

A:

The unattended flow continues after an employee leaves and fails safely on permission or data errors. This step belongs to Dataverse connector service principal and retry policies rather than the complementary pair here. The required action is: Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution. And its evidence is: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

C:

Inspect output parameters, stage, post image configuration, async mode, transaction state and external dependencies. This step belongs to Dataverse PostOperation plug-in and after-save evidence rather than the complementary pair here. The required action is: Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution. And its evidence is: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

D:

Define a Dataverse custom API with request/response properties, allowed caller, binding type and registered plug-in implementation. This step belongs to Dataverse custom API contract and plug-in execution rather than the complementary pair here. The required action is: Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution. And its evidence is: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

Question 174

A team registers an unauthenticated public webhook URL and trusts any caller who includes a contact ID. Which evidence should the developer inspect FIRST to distinguish the root cause or verify the prerequisite?

  1. Inspect registration tool endpoint URL, authentication method, execution mode, delivery logs and returned status codes.
  2. Inspect key attributes, target Entity, UpsertResponse RecordCreated, error handling, ownership and retry idempotency.
  3. Inspect Function authentication, request/response mapping, OpenAPI action, error status, timeout and log correlation.
  4. Inspect delivery attempt logs, event ID, consumer receipt store, downstream transaction and retry/backoff handling.

Explanation

Correct Answer

A:

Inspect registration tool endpoint URL, authentication method, execution mode, delivery logs and returned status codes. These observations separate this failure from adjacent capabilities or permissions. A webhook endpoint is an external receiver, and its response does not guarantee later business processing succeeded. The relevant implementation is: Register a supported Dataverse webhook service endpoint with secure HTTP endpoint and a verified listener contract.

Incorrect Answers

B:

Inspect key attributes, target Entity, UpsertResponse RecordCreated, error handling, ownership and retry idempotency. These checks are relevant to Dataverse UpsertRequest idempotent load, but are not the most discriminating first evidence for this case. Inspect instead: Inspect registration tool endpoint URL, authentication method, execution mode, delivery logs and returned status codes.

C:

Inspect Function authentication, request/response mapping, OpenAPI action, error status, timeout and log correlation. These checks are relevant to Azure Function-backed connector transformation, but are not the most discriminating first evidence for this case. Inspect instead: Inspect registration tool endpoint URL, authentication method, execution mode, delivery logs and returned status codes.

D:

Inspect delivery attempt logs, event ID, consumer receipt store, downstream transaction and retry/backoff handling. These checks are relevant to Dataverse event endpoint deduplication and retry, but are not the most discriminating first evidence for this case. Inspect instead: Inspect registration tool endpoint URL, authentication method, execution mode, delivery logs and returned status codes.

Question 175

Two registrations of the same Create handler run on one record and duplicate external notifications. Which evidence should the developer inspect FIRST to distinguish the root cause or verify the prerequisite?

  1. Inspect HTTP response code, Retry-After header, request volume, concurrency, idempotency key and retry budget.
  2. Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.
  3. Check identity registration, target resource audience, cloud tenant, scoped service role, event contract and telemetry.
  4. Inspect evaluator cases, tool traces, policy refusal tests, human approval status, error rates and rollback evidence.

Explanation

Correct Answer

B:

Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer. These observations separate this failure from adjacent capabilities or permissions. Publishing a DLL without a registered step does not automatically attach code to Dataverse messages. The relevant implementation is: Register a signed plug-in assembly, target message, step mode/stage and image configuration with supported tooling and deploy it in a solution.

Incorrect Answers

A:

Inspect HTTP response code, Retry-After header, request volume, concurrency, idempotency key and retry budget. These checks are relevant to Dataverse API service-protection throttle retry, but are not the most discriminating first evidence for this case. Inspect instead: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

C:

Check identity registration, target resource audience, cloud tenant, scoped service role, event contract and telemetry. These checks are relevant to Plug-in managed identity and business events, but are not the most discriminating first evidence for this case. Inspect instead: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

D:

Inspect evaluator cases, tool traces, policy refusal tests, human approval status, error rates and rollback evidence. These checks are relevant to Agent evaluation and human approval for side effects, but are not the most discriminating first evidence for this case. Inspect instead: Inspect assembly version, registered type, message filters, sandbox mode, image alias and effective solution layer.

Question 176

A code app intermittently times out in production but emits only a daily active-users dashboard. Which observable test result BEST confirms that the intended Power Platform component behaves correctly?

  1. Users launch the intended app revision and access only assigned data; a rollback is available.
  2. A component propagates intentional changes once and removes handlers/timers when destroyed or navigated away.
  3. The intended app instance uses the new control after import without manually copying source files.
  4. A failed operation can be traced from frontend error to server response without logging secrets or personal data.

Explanation

Correct Answer

D:

A failed operation can be traced from frontend error to server response without logging secrets or personal data. This is an observable application, data or security outcome, not merely a successful build message. Telemetry needs context and privacy controls; a page-view count cannot diagnose specific failed API calls.

Incorrect Answers

A:

Users launch the intended app revision and access only assigned data; a rollback is available. This result validates Code app packaging sharing and deployment instead of the component in question. Here the specific acceptance check is: A failed operation can be traced from frontend error to server response without logging secrets or personal data.

B:

A component propagates intentional changes once and removes handlers/timers when destroyed or navigated away. This result validates PCF lifecycle methods and output notification instead of the component in question. Here the specific acceptance check is: A failed operation can be traced from frontend error to server response without logging secrets or personal data.

C:

The intended app instance uses the new control after import without manually copying source files. This result validates Packaging and deploying PCF in solution instead of the component in question. Here the specific acceptance check is: A failed operation can be traced from frontend error to server response without logging secrets or personal data.

Question 177

The Function logs success before writing the downstream record, preventing recovery of a partly completed operation. Which TWO complementary steps provide the intended implementation AND the evidence needed to validate it? Select TWO.

  1. Inspect asynchronous request state, queue work item ID, job timeout, concurrency limits and idempotent completion marker.
  2. Use delegated OAuth for a signed-in user's work and client credentials with a constrained application user for unattended integration.
  3. Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events.
  4. The handler reliably detects a value transition even when the incoming Update contains only changed columns.
  5. Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs.

Explanation

Correct Answers

C:

Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events. This choice implements the controlling platform requirement rather than masking the symptom. At-least-once messages and function retries require idempotency at the actual business operation. Verify the change using: Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs. This is the implementation half of the required two-step answer.

E:

Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs. These observations separate this failure from adjacent capabilities or permissions. At-least-once messages and function retries require idempotency at the actual business operation. The relevant implementation is: Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events. This is the diagnostic/verification half of the required two-step answer.

Incorrect Answers

A:

Inspect asynchronous request state, queue work item ID, job timeout, concurrency limits and idempotent completion marker. This step belongs to Asynchronous background operations and safe batching rather than the complementary pair here. The required action is: Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events. And its evidence is: Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs.

B:

Use delegated OAuth for a signed-in user's work and client credentials with a constrained application user for unattended integration. This step belongs to OAuth delegated vs client credential in Dataverse rather than the complementary pair here. The required action is: Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events. And its evidence is: Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs.

D:

The handler reliably detects a value transition even when the incoming Update contains only changed columns. This step belongs to Plug-in preimage postimage and filtering attributes rather than the complementary pair here. The required action is: Use a stable correlation/business key, durable receipt and bounded retry logic in Azure Functions consuming Power Platform events. And its evidence is: Inspect event ID, durable dedup store, processing status, retry count, poison-message handling and reconciliation logs.

Question 178

An Update plug-in fires for every note-field edit because no filtering attributes were configured. Which evidence should the developer inspect FIRST to distinguish the root cause or verify the prerequisite?

  1. Inspect agent endpoint, correlation, delegated claims, response schema, tool approval and timeout behavior.
  2. Review registered image aliases, selected image columns, Target presence, MessageName and filtering attributes.
  3. Inspect output parameters, stage, post image configuration, async mode, transaction state and external dependencies.
  4. Inspect HTTP response code, Retry-After header, request volume, concurrency, idempotency key and retry budget.

Explanation

Correct Answer

B:

Review registered image aliases, selected image columns, Target presence, MessageName and filtering attributes. These observations separate this failure from adjacent capabilities or permissions. A Target entity can be sparse and an image only contains registered attributes; images do not automatically include all fields. The relevant implementation is: Register required pre/post entity images and filtering attributes to compare only relevant changes without an extra retrieve.

Incorrect Answers

A:

Inspect agent endpoint, correlation, delegated claims, response schema, tool approval and timeout behavior. These checks are relevant to Foundry to Copilot Studio agent orchestration, but are not the most discriminating first evidence for this case. Inspect instead: Review registered image aliases, selected image columns, Target presence, MessageName and filtering attributes.

C:

Inspect output parameters, stage, post image configuration, async mode, transaction state and external dependencies. These checks are relevant to Dataverse PostOperation plug-in and after-save evidence, but are not the most discriminating first evidence for this case. Inspect instead: Review registered image aliases, selected image columns, Target presence, MessageName and filtering attributes.

D:

Inspect HTTP response code, Retry-After header, request volume, concurrency, idempotency key and retry budget. These checks are relevant to Dataverse API service-protection throttle retry, but are not the most discriminating first evidence for this case. Inspect instead: Review registered image aliases, selected image columns, Target presence, MessageName and filtering attributes.

Question 179

A developer adds a new numeric control property in code but never adds the required manifest declaration. Which Power Platform extension or architecture change BEST addresses the core requirement?

  1. Configure allowed origins for scripts, styles, connections and frames according to the code app CSP rather than setting a permissive wildcard.
  2. Declare bound/input properties and resource files correctly in ControlManifest.Input.xml and align TypeScript output shapes with metadata.
  3. Package PCF through supported tooling, include built assets in a solution and bind the component to the intended column or dataset.
  4. Invoke a solution-managed cloud flow from a code app with well-defined input/output schemas and bounded timeout/error handling.

Explanation

Correct Answer

B:

Declare bound/input properties and resource files correctly in ControlManifest.Input.xml and align TypeScript output shapes with metadata. This choice implements the controlling platform requirement rather than masking the symptom. PCF manifests are contracts for host binding; TypeScript property names alone do not register fields. Verify the change using: Inspect control manifest property of-type, usage, component namespace, platform libraries and compiled resource paths.

Incorrect Answers

A:

Configure allowed origins for scripts, styles, connections and frames according to the code app CSP rather than setting a permissive wildcard. That action addresses Code app Content Security Policy configuration instead. A CSP error cannot be reliably fixed by disabling every restriction; precise origins should be justified. For this case, the targeted action is: Declare bound/input properties and resource files correctly in ControlManifest.Input.xml and align TypeScript output shapes with metadata.

C:

Package PCF through supported tooling, include built assets in a solution and bind the component to the intended column or dataset. That action addresses Packaging and deploying PCF in solution instead. A successful npm build is not equivalent to deploying and configuring a PCF in the target environment. For this case, the targeted action is: Declare bound/input properties and resource files correctly in ControlManifest.Input.xml and align TypeScript output shapes with metadata.

D:

Invoke a solution-managed cloud flow from a code app with well-defined input/output schemas and bounded timeout/error handling. That action addresses Code app use of Power Automate cloud flow instead. Calling a flow asynchronously does not guarantee its downstream actions committed before a UI success toast. For this case, the targeted action is: Declare bound/input properties and resource files correctly in ControlManifest.Input.xml and align TypeScript output shapes with metadata.

Question 180

A background integration succeeds as a system administrator but the production service principal must edit only its permitted accounts. Which observable test result BEST confirms that the intended Power Platform component behaves correctly?

  1. The service accesses only approved rows and forbidden updates fail even when made through a direct Web API call.
  2. Callers reach the intended business operation through a stable contract with its security boundary explicit.
  3. The UX extension runs in its intended host, remains maintainable and respects platform permissions.
  4. The agent executes only permitted actions and produces traceable responses under the invoking user's entitlements.

Explanation

Correct Answer

A:

The service accesses only approved rows and forbidden updates fail even when made through a direct Web API call. This is an observable application, data or security outcome, not merely a successful build message. UI hiding and a valid OAuth token alone do not establish Dataverse record authorization.

Incorrect Answers

B:

Callers reach the intended business operation through a stable contract with its security boundary explicit. This result validates Custom connector boundary versus plug-in endpoint instead of the component in question. Here the specific acceptance check is: The service accesses only approved rows and forbidden updates fail even when made through a direct Web API call.

C:

The UX extension runs in its intended host, remains maintainable and respects platform permissions. This result validates Choosing client scripts PCF or code apps instead of the component in question. Here the specific acceptance check is: The service accesses only approved rows and forbidden updates fail even when made through a direct Web API call.

D:

The agent executes only permitted actions and produces traceable responses under the invoking user's entitlements. This result validates Agent integration architecture using Copilot Studio instead of the component in question. Here the specific acceptance check is: The service accesses only approved rows and forbidden updates fail even when made through a direct Web API call.

Microsoft Power Platform Developer (PL-400) Practice Tests