Microsoft AB-100 (Agentic AI Business Solutions Architect) Practice Test 24

Question 230

Operations has not established the cause of this defect: Users cannot tell whether a value was generated or saved. The associated setting is: Staff reject suggestions because approval controls are unclear. Which inspection should precede remediation?

  1. Observe error-prone approval and revision actions in usability tests
  2. Review target host support and publishing settings
  3. Trace page data bindings and invoked record APIs
  4. Compare generated recommendations with recorded dispositions

Explanation

Correct Answer

A:

The investigation directly tests the causal connection: clear review interfaces surface provenance and editable proposals; it collects evidence for a workflow where staff reject suggestions because approval controls are unclear but the outcome is instead that users cannot tell whether a value was generated or saved.

Incorrect Answers

B:

For the failure that an agent package is assumed to work in every host automatically, reviewing target host support and publishing settings is relevant; this question instead reports that users cannot tell whether a value was generated or saved, so different evidence is needed.

C:

Tracing page data bindings and invoked record apis explores a separate connection, where generated interface components can be grounded in governed app data and defined actions; the diagnosed requirement is staff reject suggestions because approval controls are unclear, and the actual defect is users cannot tell whether a value was generated or saved.

D:

That probe evaluates a persistent agent activity feed supports review, assignment and disposition in a workflow where operations teams review multiple proposed next actions daily; the reported failure concerns the path for which clear review interfaces surface provenance and editable proposals instead.

Question 231

A business stakeholder describes this operating situation: A business process mixes deterministic calculations and open-ended judgments. Testing instead shows that a single agent unpredictably handles policy rules and narrative summarization. What design approach is most appropriate?

  1. Define cross-service boundaries and ownership before integration
  2. Use a deterministic Power Automate flow for stable approvals
  3. Decompose end-to-end workflows into agent-appropriate tasks
  4. Use Dataverse as the governed handoff for customer cases

Explanation

Correct Answer

C:

The mechanism that task decomposition separates deterministic services from bounded agent reasoning fits the scenario where a business process mixes deterministic calculations and open-ended judgments; it addresses the defect that a single agent unpredictably handles policy rules and narrative summarization.

Incorrect Answers

A:

For copilot, Dataverse, Dynamics and custom APIs share one process, this other approach relies on the fact that contracts and ownership boundaries identify responsibility for each state change; the issue here is a single agent unpredictably handles policy rules and narrative summarization; that mechanism does not establish the required task decomposition separates deterministic services from bounded agent reasoning.

B:

Explicit conditions execute known business rules predictably and leave an auditable branch history; its main purpose is the task follows fixed approval branches with no interpretation needed, not correction of the reported defect where a single agent unpredictably handles policy rules and narrative summarization in the target process.

D:

Shared authorized records support coordinated business context; its main purpose is sales and service agents need the same customer and case context, not correction of the reported defect where a single agent unpredictably handles policy rules and narrative summarization in the target process.

Question 232

One enterprise rollout has 2 separately testable requirements. (1) Business use: Complaints use vague labels such as agent is wrong. Known defect: Every issue is routed to prompt engineers despite diverse causes. (2) Business use: Users report stale answers after a policy refresh. Known defect: Feedback tickets lack the evidence needed to reproduce them. Which 2 diagnostic procedures directly satisfy these stated cases? Choose TWO.

  1. Sample feedback and inspect whether root cause categories are actionable
  2. Compare complaint time with retrieved source and model versions
  3. Review threshold breaches against user-impacting failure times
  4. Inspect clustered examples for shared underlying cause

Explanation

Correct Answers

A:

Sampling feedback and inspecting whether root cause categories are actionable can reveal why every issue is routed to prompt engineers despite diverse causes when complaints use vague labels such as agent is wrong; that finding is needed alongside the independent diagnosis for users report stale answers after a policy refresh.

B:

Comparing complaint time with retrieved source and model versions can reveal why feedback tickets lack the evidence needed to reproduce them when users report stale answers after a policy refresh; that finding is needed alongside the independent diagnosis for complaints use vague labels such as agent is wrong.

Incorrect Answers

C:

This diagnosis is associated with the different failure where teams discover failures through complaints several days later in a workflow where operations must distinguish healthy traffic from quality incidents; it does not directly solve the listed conditions: every issue is routed to prompt engineers despite diverse causes; feedback tickets lack the evidence needed to reproduce them.

D:

This diagnosis is associated with the different failure where analysts overlook a widespread intent failure among unique phrasings in a workflow where thousands of transcripts contain recurring misunderstandings; it does not directly solve the listed conditions: every issue is routed to prompt engineers despite diverse causes; feedback tickets lack the evidence needed to reproduce them.

Question 233

A test plan focuses on a previously missed failure: The agent confidently answers when retrieval fails. It occurred when policy answers must be tied to a current source. Which check directly validates that behavior?

  1. Evaluate a representative multilingual holdout
  2. Disable knowledge access and verify safe no-evidence behavior
  3. Reproduce a reported answer from recorded non-sensitive context
  4. Submit urgent and ambiguous cases to the candidate agent

Explanation

Correct Answer

B:

The scenario is that policy answers must be tied to a current source; the defect to rule out is that the agent confidently answers when retrieval fails; testing via disabling knowledge access and verifying safe no-evidence behavior challenges the design dependency where evidence and refusal evaluations enforce source-supported claims.

Incorrect Answers

A:

This check exercises a different system behavior: segmented evaluations reveal uneven model behavior; a pass there cannot establish that the agent confidently answers when retrieval fails no longer occurs in the required workflow.

C:

Although the exercise might catch feedback tickets lack the evidence needed to reproduce them, it would not directly validate the mechanism that evidence and refusal evaluations enforce source-supported claims in the target implementation.

D:

The tested behavior is that controlled service evaluations validate routing before rollout; the necessary evidence here concerns evidence and refusal evaluations enforce source-supported claims and the previously observed the agent confidently answers when retrieval fails.

Question 234

The rollout supports this workflow: A new topic description changes which agent is invoked. Leaders want a metric that responds to the reported shortcoming that regression appears only for compound user requests. Which measure best reflects success?

  1. Incorrect tool selections identified before side effects
  2. Failed agent requests with diagnosable trace paths
  3. Plan selection agreement on a frozen regression set
  4. Tokens consumed per verified completed task
  5. Production agent components using approved workload identity

Explanation

Correct Answer

C:

Because trace comparisons reveal shifts in plan selection and step order, the outcome worth following is plan selection agreement on a frozen regression set; the metric is tied to the reported issue—regression appears only for compound user requests—for the scenario where a new topic description changes which agent is invoked.

Incorrect Answers

A:

For the situation where the planner calls an order-cancel tool for status requests, incorrect tool selections identified before side effects is a sensible indicator; the problem that regression appears only for compound user requests requires measurement of plan selection agreement on a frozen regression set.

B:

The alternative tracks failed agent requests with diagnosable trace paths, arising from the mechanism where agent trace spans provide component-level execution evidence through connected telemetry; its movement alone says little about the failure that regression appears only for compound user requests.

D:

The premise of this indicator is token telemetry isolates prompt, retrieval and generation cost changes; the architecture being evaluated instead serves a new topic description changes which agent is invoked, and its success is measured by plan selection agreement on a frozen regression set.

E:

For the situation where a backend agent must call an approved Azure service without embedded secrets, production agent components using approved workload identity is a sensible indicator; the problem that regression appears only for compound user requests requires measurement of plan selection agreement on a frozen regression set.

Question 235

The rollout supports this workflow: Business applications depend on the agent endpoint continuously. Leaders want a metric that responds to the reported shortcoming that a model upgrade breaks downstream response formatting. Which measure best reflects success?

  1. Ingested documents passing quality and access gates
  2. Foundry agent releases with reproducible manifests
  3. Production calls routed to approved production dependencies
  4. User-impacting failures during agent release changes

Explanation

Correct Answer

D:

For the use case where business applications depend on the agent endpoint continuously, the indicator user-impacting failures during agent release changes is worth baselining against the earlier issue where a model upgrade breaks downstream response formatting.

Incorrect Answers

A:

That scorecard concerns the activity in which new documents enter a production source automatically and its result of ingested documents passing quality and access gates; the reported shortcoming here is a model upgrade breaks downstream response formatting.

B:

The premise of this indicator is a complete agent release manifest pins related runtime artifacts; the architecture being evaluated instead serves business applications depend on the agent endpoint continuously, and its success is measured by user-impacting failures during agent release changes.

C:

The proposed metric concerns a flow moved to production still points to a test endpoint, whereas the failure in this case is that a model upgrade breaks downstream response formatting; the indicator needed here is user-impacting failures during agent release changes.

Question 236

The pilot needs a durable correction: A replaced model package introduces unknown behavior. This affects the workflow in which deployment automation pulls model artifacts from unverified origins. Which design choice should the architect approve?

  1. Pin approved models and dependencies in a reviewed registry
  2. Register each trained model with dataset and evaluation provenance
  3. Record immutable model deployment events with approver and version
  4. Isolate model evaluation and fine-tuning data by classification

Explanation

Correct Answer

A:

The architecture must reconcile two facts: deployment automation pulls model artifacts from unverified origins, yet a replaced model package introduces unknown behavior; the selected pattern does so through the mechanism that artifact provenance and integrity checks guard model supply chain.

Incorrect Answers

B:

A model registry ties versions to lineage, evaluations and approval; its main purpose is data scientists cannot explain why a newer classifier behaves differently, not correction of the reported defect where a replaced model package introduces unknown behavior in the target process.

C:

Deployment audit links actor, approved artifact and release time; its main purpose is a model is replaced without a traceable change ticket, not correction of the reported defect where a replaced model package introduces unknown behavior in the target process.

D:

In the case where proprietary records are copied to an open test workspace, the alternate design depends on segregated datasets and access controls protect sensitive data; it does not correct the target failure, a replaced model package introduces unknown behavior.

Question 237

A policy review covers an agent for which external credit service responds slowly during regional outages. Its most relevant exposure is that credit-dependent approvals might proceed while the live credit source is unavailable. What safeguard should be required before release?

  1. Assign glossary approval to service process owners
  2. Require reviewer approval before outbound communication
  3. Prevent automatic approvals when the credit source is unavailable
  4. Exclude sensitive attributes not needed by the receiving agent

Explanation

Correct Answer

C:

Avoiding the consequence where credit-dependent approvals might proceed while the live credit source is unavailable requires a control for external credit service responds slowly during regional outages; in this architecture, bounded timeouts and graceful fallback preserve essential user work, which is the point at which the boundary can be applied.

Incorrect Answers

A:

Another process could need this guardrail because unreviewed synonyms could redefine legally significant customer-service classifications; the design being assessed supports external credit service responds slowly during regional outages and has the explicit exposure that credit-dependent approvals might proceed while the live credit source is unavailable.

B:

Another process could need this guardrail because aI-drafted customer responses might be sent before the responsible reviewer approves; the design being assessed supports external credit service responds slowly during regional outages and has the explicit exposure that credit-dependent approvals might proceed while the live credit source is unavailable.

D:

The alternate boundary mitigates a case handoff could include personal attributes irrelevant to the receiving team's work, which is relevant to agents need structured case summaries at handoff; it leaves the current workflow vulnerable to credit-dependent approvals might proceed while the live credit source is unavailable.

Question 238

A proposed deployment supports this business case: Competitive requirements depend on unique multi-model control. Reviewers also warn that a custom build might reach production without the needed operational controls. Which restriction mitigates that concrete risk?

  1. Budget for independent hosting, safety testing and monitoring
  2. Fund security, evaluation, hosting and incident response
  3. Gate launch on ownership, runbooks and support budgets
  4. Set enforceable monetary and scope thresholds outside prompts

Explanation

Correct Answer

B:

Because a custom build might reach production without the needed operational controls, a guardrail must constrain the actual system behavior: custom engineering trades more control for more operating responsibility; protecting a response alone cannot enforce that boundary.

Incorrect Answers

A:

Where a workflow needs proprietary scheduling, model choice and runtime control, the guard can address a bespoke runtime might be shipped without an accountable security or support team; the exposure that a custom build might reach production without the needed operational controls remains relevant to the present assessment.

C:

Another process could need this guardrail because production agents could launch without an incident owner, runbook or operations funding; the design being assessed supports competitive requirements depend on unique multi-model control and has the explicit exposure that a custom build might reach production without the needed operational controls.

D:

The agent may refund large amounts or deactivate accounts; its guardrail addresses the possibility that a model could override a monetary limit through conversational instructions; the present architecture must constrain the different threat that a custom build might reach production without the needed operational controls.

Question 239

The fault affects an AI-powered workflow in which multiple AI teams consume a changing source API. Specifically, an upstream field rename silently breaks downstream grounding. Which diagnostic investigation best distinguishes the relevant failure path?

  1. Inspect action failures for bad types and entity mismatches
  2. Compare event delivery logs with receiving workflow state
  3. Audit decision state across sales and finance transactions
  4. Compare consumer errors to recent contract version changes

Explanation

Correct Answer

D:

An upstream field rename silently breaks downstream grounding; to attribute the defect rather than guess at a fix, comparing consumer errors to recent contract version changes examines the dependency where versioned schemas and compatibility policy make interfaces predictable.

Incorrect Answers

A:

The proposed probe concerns the model submits malformed identifiers to an operational API, where typed inputs and server validation reject malformed or out-of-scope calls; this incident instead requires evidence about versioned schemas and compatibility policy make interfaces predictable and the failure that an upstream field rename silently breaks downstream grounding.

B:

Invoice holds must trigger exception handling without polling; its matching diagnostic check is reasonable for delayed reconciliation hides material fulfillment blockers; the present investigation needs evidence for the different failure where an upstream field rename silently breaks downstream grounding.

C:

For the failure that a suggested discount bypasses delegated approval authority, auditing decision state across sales and finance transactions is relevant; this question instead reports that an upstream field rename silently breaks downstream grounding, so different evidence is needed.

Question 240

An enterprise team is making a platform design decision. Business context: Sales copilots should not expose sensitive margin assumptions. Present failure: A connector returns the full ERP payload including protected fields. Which approach resolves the mismatch?

  1. Restrict Sales connector fields to approved data classifications
  2. Use connector actions for authenticated data mutations
  3. Enforce identity confirmation before account-specific actions
  4. Design timeout and degraded-mode handling for Sales enrichments

Explanation

Correct Answer

A:

The defect is that a connector returns the full ERP payload including protected fields; the expected business case is that sales copilots should not expose sensitive margin assumptions; the design fits because field projection and endpoint authorization minimize exposed data.

Incorrect Answers

B:

For another type of architecture, an authorized connector action invokes an operational service is useful; this design must instead support sales copilots should not expose sensitive margin assumptions despite the problem where a connector returns the full ERP payload including protected fields.

C:

For another type of architecture, authentication and authorization gates separate general help from account control is useful; this design must instead support sales copilots should not expose sensitive margin assumptions despite the problem where a connector returns the full ERP payload including protected fields.

D:

The alternative addresses a sales workflow hangs awaiting optional enrichment by a design in which bounded timeouts and graceful fallback preserve essential user work; it does not supply the missing relationship for sales copilots should not expose sensitive margin assumptions, where a connector returns the full ERP payload including protected fields.

Agentic AI Business Solutions Architect (AB-100) Practice Tests