Microsoft AB-100 (Agentic AI Business Solutions Architect) Practice Test 116

Question 1150

Risk reviewers need a control, not merely a monitoring dashboard. The use case is: An answer contradicts the handbook despite high model fluency. Potential harm: The team might publish a model that never retrieves the evidence it requires. Which protection fits?

  1. Set cost budgets and investigate anomalous context expansion
  2. Do not share private cached outputs across principals
  3. Block publishing when required evidence cannot be recovered
  4. Fail closed for unsupported sensitive answers

Explanation

Correct Answer

C:

When the team might publish a model that never retrieves the evidence it requires, the use case an answer contradicts the handbook despite high model fluency requires a guardrail aligned with the behavior that separate relevance and groundedness tests isolate the failing layer.

Incorrect Answers

A:

That protection is tied to the mechanism that token telemetry isolates prompt, retrieval and generation cost changes; it does not prevent the distinct risk where the team might publish a model that never retrieves the evidence it requires in the system under review.

B:

Another process could need this guardrail because a user might receive another principal's cached protected answer; the design being assessed supports an answer contradicts the handbook despite high model fluency and has the explicit exposure that the team might publish a model that never retrieves the evidence it requires.

D:

The safeguard for an assistant responds fluently but cites unrelated support articles deals with a different threat: missing retrieval evidence might still generate definitive sensitive advice; even after it is enforced, the exposure that the team might publish a model that never retrieves the evidence it requires remains possible in the target workflow where separate relevance and groundedness tests isolate the failing layer.

Question 1151

An incident review has one confirmed symptom: Agent invokes the wrong booking action for a reschedule. For the use case where two topics claim nearly identical scheduling intents, which targeted diagnostic check would reveal the most pertinent evidence?

  1. Inspect component inputs and app-to-flow error handling
  2. Inspect tool permissions and response sources
  3. Review action arguments and created escalation ownership
  4. Compare competing topic traces for representative phrasing

Explanation

Correct Answer

D:

In the setting where two topics claim nearly identical scheduling intents, comparing competing topic traces for representative phrasing can isolate the failure in which agent invokes the wrong booking action for a reschedule, because distinct descriptions and inputs improve topic selection.

Incorrect Answers

A:

Inspectors submit a short description and need suggested categorization; its matching diagnostic check is reasonable for free-form AI output is written to the wrong Dataverse field; the present investigation needs evidence for the different failure where agent invokes the wrong booking action for a reschedule.

B:

The evidence this method produces applies to a workflow where marketing specialists need tailored descriptions from existing briefings; it cannot by itself establish why agent invokes the wrong booking action for a reschedule in the current scenario.

C:

The proposed inspection is associated with a representative must initiate an exception workflow from the case, where an authorized action creates a governed workflow item from the current case; it would not expose the failure mode that agent invokes the wrong booking action for a reschedule with the required specificity.

Question 1152

A risk analysis highlights a threat in this workflow: Agents can call APIs with customer IDs and amount values. Exposure: A hallucinated expected outcome might become a trusted automated oracle. Which enforceable safeguard addresses it at the relevant boundary?

  1. Avoid privileged shared accounts for user-specific retrieval
  2. Block release on any high-severity authorization regression
  3. Forbid direct mutation before recorded delegated approval
  4. Treat generated expectations as suggestions until verified
  5. Require explicit approval when critical thresholds fail

Explanation

Correct Answer

D:

Agents can call APIs with customer IDs and amount values; its risk scenario is that a hallucinated expected outcome might become a trusted automated oracle; the safeguard fits because specification-guided generation covers missing fields and denied roles and the restriction addresses that specific exposure.

Incorrect Answers

A:

The alternate boundary mitigates a shared admin token might let ordinary users retrieve executive records, which is relevant to agent looks up records on behalf of employees with different rights; it leaves the current workflow vulnerable to a hallucinated expected outcome might become a trusted automated oracle.

B:

The alternate boundary mitigates a release might ship with an exploitable record-access regression, which is relevant to agent supports refunds, order status and generic FAQs; it leaves the current workflow vulnerable to a hallucinated expected outcome might become a trusted automated oracle.

C:

This restriction is associated with service recommends a price adjustment requiring finance approval and addresses the exposure that a suggested customer concession might post before delegated finance approval; it does not specifically protect the target workflow against a hallucinated expected outcome might become a trusted automated oracle.

E:

Another process could need this guardrail because a maker might publish a candidate despite a failed critical evaluation gate; the design being assessed supports agents can call APIs with customer IDs and amount values and has the explicit exposure that a hallucinated expected outcome might become a trusted automated oracle.

Question 1153

The operating team must monitor whether an architectural correction works. Business context: Users must remain in a proprietary work-order application. Historical symptom: Requiring a separate chat portal increases context switching. Which outcome measure is the best match?

  1. Percentage of delegated tasks completed within agreed deadlines
  2. Authorized tool calls completed without excess privileges
  3. Task completion without switching application context
  4. Custom workflow completion against the required execution contract

Explanation

Correct Answer

C:

Because app-hosted integration brings relevant AI actions into an owned user interface, the outcome worth following is task completion without switching application context; the metric is tied to the reported issue—requiring a separate chat portal increases context switching—for the scenario where users must remain in a proprietary work-order application.

Incorrect Answers

A:

The proposed metric concerns downstream research agents run longer than the interactive request, whereas the failure in this case is that requiring a separate chat portal increases context switching; the indicator needed here is task completion without switching application context.

B:

A Copilot Studio agent needs a reusable interface to governed operations; its intended measurement is authorized tool calls completed without excess privileges; substituting that result here would not capture whether requiring a separate chat portal increases context switching has been corrected.

D:

The premise of this indicator is custom-engine architecture controls runtime, model calls and integrations; the architecture being evaluated instead serves users must remain in a proprietary work-order application, and its success is measured by task completion without switching application context.

Question 1154

An operational risk could arise because a new tokenizer might be deployed with an old label mapping. The AI system supports the business need that a retraining release changes the tokenizer and label order. Which guardrail is designed for this threat?

  1. Limit retroactive alteration of approved QA evidence
  2. Require mitigation before deployment when disparity is material
  3. Reject mismatched schema/model combinations
  4. Restrict promotion to reviewed compatible configurations

Explanation

Correct Answer

C:

Because a new tokenizer might be deployed with an old label mapping, a guardrail must constrain the actual system behavior: bundled preprocessing, model and label maps maintain contract consistency; protecting a response alone cannot enforce that boundary.

Incorrect Answers

A:

The safeguard for an updated agent performs differently months after release deals with a different threat: old evaluation evidence might be overwritten to conceal a failed release gate; even after it is enforced, the exposure that a new tokenizer might be deployed with an old label mapping remains possible in the target workflow where bundled preprocessing, model and label maps maintain contract consistency.

B:

Another process could need this guardrail because an unfair subgroup outcome might remain unresolved at promotion; the design being assessed supports a retraining release changes the tokenizer and label order and has the explicit exposure that a new tokenizer might be deployed with an old label mapping.

D:

Model and tool changes are promoted independently; its guardrail addresses the possibility that an agent release might pair instructions with incompatible tool or model versions; the present architecture must constrain the different threat that a new tokenizer might be deployed with an old label mapping.

Question 1155

A policy review covers an agent for which a won opportunity should create a properly owned onboarding case. Its most relevant exposure is that duplicate Dynamics events might create redundant onboarding cases. What safeguard should be required before release?

  1. Separate draft and send permissions in service integrations
  2. Block release if privileged writes bypass approvals
  3. Prevent duplicate case creation on repeated event delivery
  4. Prevent actions before identity and missing fields are verified

Explanation

Correct Answer

C:

Because duplicate Dynamics events might create redundant onboarding cases, a guardrail must constrain the actual system behavior: end-to-end tests verify identity, mapping and assignment across apps; protecting a response alone cannot enforce that boundary.

Incorrect Answers

A:

Another process could need this guardrail because a service drafting assistant might gain permission to send unapproved notices; the design being assessed supports a won opportunity should create a properly owned onboarding case and has the explicit exposure that duplicate Dynamics events might create redundant onboarding cases.

B:

That protection is tied to the mechanism that authorization regression gates verify posting authority remains constrained; it does not prevent the distinct risk where duplicate Dynamics events might create redundant onboarding cases in the system under review.

D:

The alternate guard addresses an agent might act before collecting mandatory identity and form fields for agent must remember order context while asking for missing details; the present threat, duplicate Dynamics events might create redundant onboarding cases, needs a different enforcement boundary.

Question 1156

During integration testing, the agent invents discretionary promises. The associated business context is: Compliance letters must preserve approved language and disclaimers. Which architectural remedy is the strongest match?

  1. Use channel-specific accessibility and response policies
  2. Choose bounded task-specific instructions over open-ended autonomy
  3. Provide source-linked explanations for analytical responses
  4. Control creativity for regulated response templates

Explanation

Correct Answer

D:

The mechanism that templating and approved clauses limit generated variation fits the scenario where compliance letters must preserve approved language and disclaimers; it addresses the defect that the agent invents discretionary promises.

Incorrect Answers

A:

The technical behavior of this alternative is that channel-aware response design adapts wording and modality; the user case being assessed is compliance letters must preserve approved language and disclaimers with a different failure: the agent invents discretionary promises.

B:

For another type of architecture, limited intent and tool scope prevent unnecessary actions is useful; this design must instead support compliance letters must preserve approved language and disclaimers despite the problem where the agent invents discretionary promises.

C:

In the case where finance asks why a variance was flagged, the alternate design depends on grounded explanatory responses distinguish evidence from hypothesis; it does not correct the target failure, the agent invents discretionary promises.

Question 1157

Risk reviewers need a control, not merely a monitoring dashboard. The use case is: Users ask about governed CRM records rather than prose manuals. Potential harm: Dataverse retrieval might reveal a row not authorized to the caller. Which protection fits?

  1. Suppress assertions that require evidence when no source was retrieved
  2. Require idempotency at cross-service mutation boundaries
  3. Preserve table and row access controls for retrieval
  4. Quarantine unclassified or ownerless content pending review

Explanation

Correct Answer

C:

Because Dataverse retrieval might reveal a row not authorized to the caller, a guardrail must constrain the actual system behavior: Dataverse grounding exposes appropriate structured records to the agent; protecting a response alone cannot enforce that boundary.

Incorrect Answers

A:

The safeguard for required knowledge stores are intermittently unreachable deals with a different threat: a retrieval outage could make the model invent unsupported factual claims; even after it is enforced, the exposure that Dataverse retrieval might reveal a row not authorized to the caller remains possible in the target workflow where Dataverse grounding exposes appropriate structured records to the agent.

B:

This restriction is associated with copilot, Dataverse, Dynamics and custom APIs share one process and addresses the exposure that network retries could create duplicated writes across independent systems; it does not specifically protect the target workflow against Dataverse retrieval might reveal a row not authorized to the caller.

D:

Search spans internal manuals with different owners and permissions; its guardrail addresses the possibility that ownerless confidential drafts could become searchable before classification; the present architecture must constrain the different threat that Dataverse retrieval might reveal a row not authorized to the caller.

Question 1158

During integration testing, an agent package is assumed to work in every host automatically. The associated business context is: Agent experience must appear in selected Microsoft 365 contexts. Which architectural remedy is the strongest match?

  1. Extend service assistance with an approved escalation action
  2. Plan distribution separately for Teams and SharePoint surfaces
  3. Add a plugin action to a supported declarative agent
  4. Design timeout and degraded-mode handling for Sales enrichments

Explanation

Correct Answer

B:

For the need that agent experience must appear in selected Microsoft 365 contexts, the relevant mechanism is that host-specific support, packaging and admin approval shape deployment; that architectural dependency addresses the observed failure where an agent package is assumed to work in every host automatically.

Incorrect Answers

A:

An authorized action creates a governed workflow item from the current case; its main purpose is a representative must initiate an exception workflow from the case, not correction of the reported defect where an agent package is assumed to work in every host automatically in the target process.

C:

There is a valid place for the mechanism that a plugin provides authorized callable tools to an agent, namely a Teams user should query and update a supported external service; it does not remove the specific defect that an agent package is assumed to work in every host automatically from this case.

D:

A sales workflow hangs awaiting optional enrichment; the alternate design can address that situation because bounded timeouts and graceful fallback preserve essential user work, whereas this question needs a solution to an agent package is assumed to work in every host automatically.

Question 1159

A design adds autonomous behavior to a process where an agent needs to update a governed ticket status. This creates a documented risk: A tool might update tickets using a shared identity with too many permissions. Which enforced control is the best fit?

  1. Apply tenant policies and scoped content permissions
  2. Limit creation privileges and prevent duplicate escalations
  3. Disable write tools and label generated material for review
  4. Require least-privilege service permissions and action confirmation
  5. Label unsupported explanations as hypotheses or abstain
  6. Require confirmation when intent remains ambiguous

Explanation

Correct Answer

D:

Avoiding the consequence where a tool might update tickets using a shared identity with too many permissions requires a control for an agent needs to update a governed ticket status; in this architecture, an authorized connector action invokes an operational service, which is the point at which the boundary can be applied.

Incorrect Answers

A:

The proposed control targets a declarative agent might be published beyond permitted tenant groups, while the given risk scenario is a tool might update tickets using a shared identity with too many permissions; these exposures arise in different operational contexts.

B:

Where a representative must initiate an exception workflow from the case, the guard can address repeated agent calls might create duplicate escalation records or excess privileges; the exposure that a tool might update tickets using a shared identity with too many permissions remains relevant to the present assessment.

C:

This restriction is associated with marketing specialists need tailored descriptions from existing briefings and addresses the exposure that a text-only drafting assistant might be given unnecessary record-writing powers; it does not specifically protect the target workflow against a tool might update tickets using a shared identity with too many permissions.

E:

Finance asks why a variance was flagged; its guardrail addresses the possibility that an analytical response might present an unsupported hypothesis as a ledger fact; the present architecture must constrain the different threat that a tool might update tickets using a shared identity with too many permissions.

F:

For the use case where two topics claim nearly identical scheduling intents, the control addresses the threat that overlapping scheduling intents might invoke a change before confirmation; the specified agent must be protected against a tool might update tickets using a shared identity with too many permissions.

Question 1160

One enterprise rollout has 2 separately testable requirements. (1) Business use: Personal read-only assistants and payment agents have different risk. Exposure: A privileged autonomous agent might receive the same review as a private helper. (2) Business use: Business units create agents with no central visibility. Exposure: Production agents might have no recorded accountable owner or risk classification. Which 2 safeguards directly satisfy these stated cases? Choose TWO.

  1. Require heightened review for autonomous privileged agents
  2. Restore prior version on any critical schema or safety failure
  3. Block publication without accountable owner and classification
  4. Require reviewed datasets and independent promotion sign-off

Explanation

Correct Answers

A:

For the exposure that a privileged autonomous agent might receive the same review as a private helper, zoned governance aligns controls with impact and data sensitivity explains the protection required when personal read-only assistants and payment agents have different risk; this does not replace control of production agents might have no recorded accountable owner or risk classification.

C:

For the exposure that production agents might have no recorded accountable owner or risk classification, inventory metadata records owner, risk tier, tools and lifecycle state explains the protection required when business units create agents with no central visibility; this does not replace control of a privileged autonomous agent might receive the same review as a private helper.

Incorrect Answers

B:

This security safeguard addresses the unrelated exposure that a canary model might cause critical schema or safety regressions without quick recovery in a workflow where business applications depend on the agent endpoint continuously; the question instead names the specific risks a privileged autonomous agent might receive the same review as a private helper; production agents might have no recorded accountable owner or risk classification; this choice does not directly mitigate those specific exposures.

D:

This security safeguard addresses the unrelated exposure that unreviewed training examples might poison a promoted model in a workflow where business analysts can submit examples but not change production behavior; the question instead names the specific risks a privileged autonomous agent might receive the same review as a private helper; production agents might have no recorded accountable owner or risk classification; this choice does not directly mitigate those specific exposures.

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