Microsoft AB-100 Practice Test 35

Question 340

A marketplace rollback restores an earlier agent, but the refund tool retained its newer incompatible argument schema. What is the safest recovery plan?

Issue refunds through the incompatible tool and inspect the ledger next week.
Rollback the tool contract or supply a tested compatibility adapter, then verify single authorized refunds in a controlled smoke test.
Change only the agent’s visible name to the old version.
Ask the model to ignore schema errors when explaining the refund outcome.

Explanation

Correct Answer

Agent version and tool schema are coupled dependencies; rollback is safe only when the contract is compatible and verified.

Incorrect Answers

An untested incompatible tool can write wrong financial values before delayed detection.

The display name does not change arguments sent to a remote API.

Ignoring errors in prose would falsely confirm transactions.

Question 341

A global finance assistant uses a pricing workbook indexed nightly. An audit must reconstruct advice issued before a midweek correction to the workbook. Which retention choice supports this?

Keep only the latest corrected workbook to avoid retaining obsolete versions.
Retain the user’s country but not the source or model configuration.
Store a screenshot of the assistant’s final answer with no retrieved evidence IDs.
Retain immutable nightly source and index digests, the effective timestamps, and the prompt/model versions used for that request.

Explanation

Correct Answer

Effective-dated source snapshots establish what evidence was available for the historical decision at the time of issuance.

Incorrect Answers

The corrected workbook can be right today but cannot reproduce what was retrieved previously.

Country metadata does not identify a specific source revision.

A screenshot cannot substantiate the input evidence or calculation path.

Question 342

A retail Copilot Studio team wants emergency rollback without enabling citizen makers to deploy arbitrary changes to production. Which ALM measure helps?

Allow the agent to trigger deployment when it identifies its own prior version without an independent pipeline role or approval record.
Let all development makers edit production directly because it is faster than a pipeline.
Skip managed solution versioning and rely on chat history as the restore artifact.
Control solution promotion through approved pipelines and roles, and keep the previous managed release available.

Explanation

Correct Answer

Separation of duties and versioned releases protect both fast rollback and production governance.

Incorrect Answers

A model’s recognition of version metadata cannot substitute for a controlled release identity, authorization and reproducible rollback artifact.

Direct maker edits defeat audited change approval and increase accidental production drift.

Chat history is not a reproducible deployment artifact.

Question 343

A construction cost agent deployed through Power Platform has a valid production connection reference, but uses the wrong budget environment variable. Which evidence should release validation record?

The number of environments listed in the administrator portal.
A sample conversation that contains the word “budget” without calling the tool.
The resolved production budget data-source identifier and a controlled query proving the intended project and tenant are accessed.
Only the connection reference display label and a successful login.

Explanation

Correct Answer

Identity binding and configuration values are separate dependencies; test which real dataset the action uses.

Incorrect Answers

Environment count has no relationship to the resolved project identifier.

Topic words in generated text do not demonstrate actual data retrieval.

Authentication success can occur against the wrong budget source.

Question 344

An energy company’s Foundry agent candidate appears healthy in a canary but rarely invokes the emergency shutoff tool. What staged release evidence must be added?

A check that the shutoff tool appears in the agent’s list of registered tools.
A labeled safety-case suite checking correct shutoff authorization, expected tool call or refusal and resulting equipment state before expanding traffic.
Only the average chat response time across casual non-emergency requests.
Count successful emergency-shutoff explanations without executing or safely simulating the protected tool-authorization path.

Explanation

Correct Answer

Rare high-impact actions require targeted behavior and authoritative approval tests, not only broad production health metrics.

Incorrect Answers

Tool registration alone does not establish correct invocation, permissions or consequences.

Average response time can look healthy while emergency behavior is unsafe.

Verbal safety claims are not evidence that the right tool executes or refuses under the correct identity and physical state.

Question 345

A marketplace builds a second custom fraud model and must compare it against the approved production model. Which release artifact makes comparison valid?

Version both models and training datasets, evaluate on the same held-out fraud cohorts and retain per-segment error measurements.
Report each model’s accuracy on a different historical dataset with unrelated class distributions.
Compare model-generated fraud-risk explanations instead of false-positive/false-negative outcomes on common labeled cohorts.
Promote whichever model was trained most recently regardless of false-positive cost.

Explanation

Correct Answer

Controlled comparisons require common evaluation populations and traceable artifacts to attribute performance changes.

Incorrect Answers

Different datasets confound the effect of model quality with class composition.

Fluent risk explanations are not detection evidence. Comparable labeled cases and segment errors are required for a valid candidate comparison.

Newer models can introduce costly false positives or unequal subgroup errors.

Question 346

A global finance team changes Dynamics 365 tax configuration and agent approval prompts on the same day. An audit later finds wrong tax totals. What release practice would have improved diagnosis?

Version and correlate tax-configuration, application extension and agent prompt releases with transaction-level trace IDs.
Store only the final deployment release notes and omit transaction-level correlation between tax rules and agent versions.
Disable tax-rule logging during the cutover to reduce production traffic.
Deploy both changes without timing records so their owners cannot influence each other.

Explanation

Correct Answer

Cross-component release provenance and correlated transactions identify which change affected the tax calculation.

Incorrect Answers

Release notes list planned changes but do not prove which configuration was active for a disputed invoice; link exact deployed versions and transaction traces.

Removing control logs undermines accounting reconstruction.

Missing change timelines make causal diagnosis harder, not safer.

Question 347

A call-center organization updates a Customer Service agent and its Dynamics 365 Sales contact sync connector. What compatibility assurance should precede production routing?

Assume a connector health response means all mapped contact records are correct.
Verify contact ID mapping, field permissions, duplicate handling and synchronization receipts against the actual updated connector contract.
Approve because the agent can answer a generic FAQ about product pricing.
Use an old contact export without testing whether newly required fields are present.

Explanation

Correct Answer

Customer identity and consent fields must survive schema changes and be checked as persisted data across systems.

Incorrect Answers

A successful endpoint response cannot prove semantic correctness of imported records.

FAQ accuracy exercises a different path than contact synchronization.

Historical exports may not represent the new contract or required attributes.

Question 348

A construction cost-approval agent has just been rolled back after a faulty production update. Select TWO activities that belong in the immediate postrollback smoke and reconciliation plan.

Assume a displayed old version label proves all data and dependencies have been restored.
Delete all change logs created during the failed deployment before auditors inspect them.
Switch back to the faulty candidate if its answer style is more concise.
Inspect recent cost approvals and confirm any incorrect posted changes were reversed or escalated through authorized remediation.
Confirm the restored agent and its tool connection resolve to the compatible approved versions.

Explanation

Correct Answers

Rollback prevents further faults but does not repair previously committed business transactions; affected records need explicit reconciliation.

Dependencies such as tool bindings and source versions must be restored together with agent instructions for behavior to be reliable.

Incorrect Answers

Version display text alone cannot prove runtime configuration or data state.

Audit logs are needed to trace and repair the incident, not to be erased.

Answer style is irrelevant when the candidate remains known to authorize wrong transactions.

Question 349

An energy utility’s restoration agent can read outage tickets but is configured with the same identity as a control-system operator. Which security boundary is needed before production use?

Give the agent only outage-read capabilities and require a separately authorized operator workflow for any control-system command.
Hide the equipment-control tool from the agent interface while keeping its backend permission active.
Allow control commands whenever a generated explanation references a safety procedure.
Use the same production operator credential to simplify ticket retrieval and equipment control.

Explanation

Correct Answer

The agent’s available actions must reflect least privilege; a language-model instruction cannot replace separate authorization for physical effects.

Incorrect Answers

UI hiding does not revoke tool execution rights at the service boundary.

Citing a procedure does not grant permission to operate equipment.

Shared high-privilege credentials erase the boundary between observation and physical actuation.

Question 350

A marketplace has forty active autonomous agents but nobody can identify the owner of a refund-adjustment agent created last year. What governance control should be established?

A governed agent inventory with owner, purpose, approved tools, risk tier, release version and periodic access recertification.
A list of agent display names with no owner or permitted actions.
An assumption that the creator remains accountable after leaving the organization.
A single unrestricted admin credential shared by every agent to reduce inventory overhead.

Explanation

Correct Answer

Accountability depends on identifiable owners and deployed capabilities throughout the agent lifecycle, including offboarding and recertification.

Incorrect Answers

Display names alone cannot identify responsibility or high-risk tool access.

Departed employees cannot provide an enforceable ongoing governance approval.

Shared administrator identity increases risk and defeats attribution rather than simplifying governance safely.

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

Microsoft AB-100 Practice Test 33 - 10 Questions (Last Reviewed 12-Oct-2026)
Microsoft AB-100 Practice Test 34 - 10 Questions (Last Reviewed 11-Oct-2026)
Microsoft AB-100 Practice Test 35 - 10 Questions (Last Reviewed 10-Oct-2026)
Microsoft AB-100 Practice Test 36 - 10 Questions (Last Reviewed 09-Oct-2026)
Microsoft AB-100 Practice Test 37 - 10 Questions (Last Reviewed 08-Oct-2026)

Pass Your Microsoft AB-100 Exam Easy!

Question 340

Explanation

Correct Answer

Incorrect Answers

Question 341

Explanation

Correct Answer

Incorrect Answers

Question 342

Explanation

Correct Answer

Incorrect Answers

Question 343

Explanation

Correct Answer

Incorrect Answers

Question 344

Explanation

Correct Answer

Incorrect Answers

Question 345

Explanation

Correct Answer

Incorrect Answers

Question 346

Explanation

Correct Answer

Incorrect Answers

Question 347

Explanation

Correct Answer

Incorrect Answers

Question 348

Explanation

Correct Answers

Incorrect Answers

Question 349

Explanation

Correct Answer

Incorrect Answers

Question 350

Explanation

Correct Answer

Incorrect Answers

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