Question 860
The architecture board has evidence that no log identifies which source version was indexed. The affected situation is: Agents begin citing an obsolete article after an ingestion job. What change most directly establishes the missing behavior?
- Preserve reproducible evaluation evidence for major model changes
- Restrict who can supply tuning data and approve model updates
- Track grounding-source updates and index rebuild provenance
- Retain rollback snapshots for models and data transformations
Explanation
Correct Answer
C:The mechanism that content lineage records update, transformation and index revision fits the scenario where agents begin citing an obsolete article after an ingestion job; it addresses the defect that no log identifies which source version was indexed.
Incorrect Answers
A:When an updated agent performs differently months after release, versioned evaluation sets, model IDs and thresholds document release claims; in the stated workflow, however, no log identifies which source version was indexed; the relevant architectural behavior is that content lineage records update, transformation and index revision.
B:For another type of architecture, separated data-curation and release permissions preserve integrity is useful; this design must instead support agents begin citing an obsolete article after an ingestion job despite the problem where no log identifies which source version was indexed.
D:The alternative addresses restoring code alone does not restore previous predictions by a design in which versioned data transformations and model artifacts enable consistent rollback; it does not supply the missing relationship for agents begin citing an obsolete article after an ingestion job, where no log identifies which source version was indexed.
Question 861
An operational risk could arise because a privileged agent mutation might lack the initiating user and approval trail. The AI system supports the business need that a tool deletes a record following an ambiguous conversation. Which guardrail is designed for this threat?
- Reject missing, invalid or out-of-scope authorization tokens
- Reject unregistered or unapproved models from production
- Do not approve procurement while financial hold status is unknown
- Keep append-only records and redact unnecessary personal details
Explanation
Correct Answer
D:Avoiding the consequence where a privileged agent mutation might lack the initiating user and approval trail requires a control for a tool deletes a record following an ambiguous conversation; in this architecture, end-to-end audit binds user, tool, affected record and authorization event, which is the point at which the boundary can be applied.
Incorrect Answers
A:Prompt wording persuades the agent to submit a privileged operation; its guardrail addresses the possibility that an API might trust model-generated identity claims instead of valid authorization; the present architecture must constrain the different threat that a privileged agent mutation might lack the initiating user and approval trail.
B:This restriction is associated with data scientists cannot explain why a newer classifier behaves differently and addresses the exposure that a model outside the approved registry might reach production; it does not specifically protect the target workflow against a privileged agent mutation might lack the initiating user and approval trail.
C:Another process could need this guardrail because a procurement approval might proceed before a finance hold is resolved; the design being assessed supports a tool deletes a record following an ambiguous conversation and has the explicit exposure that a privileged agent mutation might lack the initiating user and approval trail.
Question 862
The rollout supports this workflow: Users need familiar Copilot behavior plus an internal status lookup. Leaders want a metric that responds to the reported shortcoming that a custom replacement duplicates the host experience unnecessarily. Which measure best reflects success?
- Outbound notices released with recorded approval
- Relevant external content found without access violations
- Successful custom action calls from the prebuilt experience
- Cross-app case-state agreement rate
Explanation
Correct Answer
C:The observed shortcoming is that a custom replacement duplicates the host experience unnecessarily; measuring successful custom action calls from the prebuilt experience gives business owners a directly relevant indicator for the process where users need familiar Copilot behavior plus an internal status lookup.
Incorrect Answers
A:A change in outbound notices released with recorded approval is informative for standard capabilities draft customer communications but policy requires approval; it does not establish whether the current process achieves successful custom action calls from the prebuilt experience.
B:For the situation where teams want approved external records searchable in Copilot, relevant external content found without access violations is a sensible indicator; the problem that a custom replacement duplicates the host experience unnecessarily requires measurement of successful custom action calls from the prebuilt experience.
D:For the situation where sales and service agents need the same customer and case context, cross-app case-state agreement rate is a sensible indicator; the problem that a custom replacement duplicates the host experience unnecessarily requires measurement of successful custom action calls from the prebuilt experience.
Question 863
A planned AI solution must support this situation: Users need ordinary meeting summaries and follow-up drafts. Its prototype fails because a costly custom runtime replicates a mature standard capability. Which design best fits the constraint?
- Extend a prebuilt agent with an approved business action
- Adopt a supported prebuilt agent for standard productivity workflows
- Use a hybrid architecture for managed commodity plus custom logic
- Buy a supported managed capability for standard requirements
Explanation
Correct Answer
B:The architecture must reconcile two facts: users need ordinary meeting summaries and follow-up drafts, yet a costly custom runtime replicates a mature standard capability; the selected pattern does so through the mechanism that a prebuilt Microsoft 365 capability minimizes custom maintenance for commodity tasks.
Incorrect Answers
A:This design would make sense if users need familiar Copilot behavior plus an internal status lookup, particularly where a custom replacement duplicates the host experience unnecessarily; the requested architecture instead needs the dependency that a prebuilt Microsoft 365 capability minimizes custom maintenance for commodity tasks because a costly custom runtime replicates a mature standard capability.
C:In the case where standard conversation needs a proprietary scoring service, the alternate design depends on a managed host can invoke an independently governed domain API; it does not correct the target failure, a costly custom runtime replicates a mature standard capability.
D:The technical behavior of this alternative is that managed functionality shifts routine platform operations to the service; the user case being assessed is users need ordinary meeting summaries and follow-up drafts with a different failure: a costly custom runtime replicates a mature standard capability.
Question 864
Troubleshooting starts from a business complaint: Generated output occasionally omits required keys. The affected process is: A Copilot prompt feeds records into downstream automation. Which investigation is most directly tied to the reported defect?
- Compare precision and recall by field and layout type
- Inspect field-missing and type-mismatch error clusters
- Review the signed acceptance matrix and results
- Trace source text influence on tool choices
- Diff tool traces and evaluation outcomes against baseline
Explanation
Correct Answer
B:Where generated output occasionally omits required keys, the useful evidence comes from inspecting field-missing and type-mismatch error clusters; its focus matches the architectural mechanism that schema validation catches responses that violate integration contracts in the process where a Copilot prompt feeds records into downstream automation.
Incorrect Answers
A:A high aggregate score masks critical field errors; examining that other path does not determine whether schema validation catches responses that violate integration contracts, which is what distinguishes the issue in the stated use case.
C:Teams promote a custom model because it feels better; its matching diagnostic check is reasonable for candidate improves one metric while violating another; the present investigation needs evidence for the different failure where generated output occasionally omits required keys.
D:External knowledge contains embedded role-change commands; its matching diagnostic check is reasonable for the agent obeys text retrieved from a supplier document; the present investigation needs evidence for the different failure where generated output occasionally omits required keys.
E:The evidence this method produces applies to a workflow where a hosted agent gains an additional research tool; it cannot by itself establish why generated output occasionally omits required keys in the current scenario.
Question 865
Support engineers report this failure: A simple scripted topic cannot handle legitimate variants. The business context is: Procurement exceptions require comparing several policies and records. What evidence should be collected to isolate the cause?
- Inspect plugin capabilities and authentication configuration
- Compare translation errors on specialist vocabulary
- Review plans for unauthorized leaps or tool chaining errors
- Inspect field-level source and suggestion labels
Explanation
Correct Answer
C:The investigation directly tests the causal connection: reasoning can plan through evidence and tools within enforced boundaries; it collects evidence for a workflow where procurement exceptions require comparing several policies and records but the outcome is instead that a simple scripted topic cannot handle legitimate variants.
Incorrect Answers
A:In the separate scenario where a Teams user should query and update a supported external service, the same investigation could explain read-only knowledge alone cannot create a service request; it does not isolate the reported cause in a system where procurement exceptions require comparing several policies and records.
B:For the failure that a monolingual analysis misses non-English escalations, comparing translation errors on specialist vocabulary is relevant; this question instead reports that a simple scripted topic cannot handle legitimate variants, so different evidence is needed.
D:The evidence this method produces applies to a workflow where users need to distinguish verified fields from AI suggestions; it cannot by itself establish why a simple scripted topic cannot handle legitimate variants in the current scenario.
Question 866
A release must prove that full replacement increases retraining and migration risk will not recur. The application is used in this situation: Current users need one governed custom action inside their workflow. Which acceptance test most directly checks that failure mode?
- Test the custom action with existing identity and audit policy
- Replace a source document without training and verify updated answers
- Run a representative extraction holdout at production concurrency
- Query an external article using users with different permissions
Explanation
Correct Answer
A:The historical failure is that full replacement increases retraining and migration risk; for the setting where current users need one governed custom action inside their workflow, testing the custom action with existing identity and audit policy verifies the dependency where extension retains familiar user experience and adds bounded custom behavior.
Incorrect Answers
B:Although the exercise might catch a trained model remembers obsolete guidance after source changes, it would not directly validate the mechanism that extension retains familiar user experience and adds bounded custom behavior in the target implementation.
C:Although the exercise might catch a general large model exceeds per-document unit economics, it would not directly validate the mechanism that extension retains familiar user experience and adds bounded custom behavior in the target implementation.
D:The suggested test serves teams want approved external records searchable in Copilot, in which the agent lacks knowledge of non-Microsoft content; the present scenario instead concerns the reported problem where full replacement increases retraining and migration risk.
Question 867
In a business situation where a hosted agent gains an additional research tool, a previous build showed this defect: New planning behavior occasionally leaks internal notes. Which reproducible test belongs in the release gate?
- Roll out to one region and check segmented resolution metrics
- Attempt to publish agents from two risk zones
- Try to send high-impact notices without approval in staging
- Run red-team and representative workflow tests before publishing
Explanation
Correct Answer
D:New planning behavior occasionally leaks internal notes; a release test must establish that this no longer occurs; running red-team and representative workflow tests before publishing measures the specific behavior for which evaluation suites validate quality and safety before publication.
Incorrect Answers
A:The verification concerning a conversational change affects only one support geography checks cohort telemetry supports safe staged validation against global averages hide a regional escalation spike; its result says little about the observed issue where new planning behavior occasionally leaks internal notes.
B:In the alternate setting where personal read-only assistants and payment agents have different risk, the test exercises zoned governance aligns controls with impact and data sensitivity; even a pass would not show that new planning behavior occasionally leaks internal notes has been prevented.
C:A pass on that verification would speak to service agents draft customer messages but cannot send without review; it would not show that evaluation suites validate quality and safety before publication, which is needed to avoid the observed defect new planning behavior occasionally leaks internal notes.
Question 868
An enterprise team is making a platform design decision. Business context: A Copilot Studio agent needs a reusable interface to governed operations. Present failure: Every agent reimplements the same proprietary tool signatures. Which approach resolves the mismatch?
- Expose governed operational entities through Dataverse
- Use A2A for agent-to-agent collaboration across runtimes
- Use MCP for an agent's access to external tools and data
- Use a Copilot Studio conversational agent for variable natural-language requests
Explanation
Correct Answer
C:The architecture must reconcile two facts: a Copilot Studio agent needs a reusable interface to governed operations, yet every agent reimplements the same proprietary tool signatures; the selected pattern does so through the mechanism that MCP presents supported tools and resources through a common integration contract.
Incorrect Answers
A:Dataverse tables and governed relationships provide a reusable business-data surface; its main purpose is agents and apps need shared definitions of customers and work orders, not correction of the reported defect where every agent reimplements the same proprietary tool signatures in the target process.
B:An agent interoperability protocol supports explicit task delegation between agents; its main purpose is independently hosted agents must exchange tasks and results, not correction of the reported defect where every agent reimplements the same proprietary tool signatures in the target process.
D:For employees express several intents and need interactive clarification, this other approach relies on the fact that generative orchestration interprets intent and can combine authorized tools and knowledge; the issue here is every agent reimplements the same proprietary tool signatures; that mechanism does not establish the required MCP presents supported tools and resources through a common integration contract.
Question 869
Risk reviewers need a control, not merely a monitoring dashboard. The use case is: A sensitive model endpoint is reachable using a broadly shared key. Potential harm: A leaked endpoint credential might permit inference from an untrusted client. Which protection fits?
- Do not approve procurement while financial hold status is unknown
- Do not share private cached outputs across principals
- Reject missing, invalid or out-of-scope authorization tokens
- Rotate compromised keys and enforce least-privilege callers
Explanation
Correct Answer
D:In the context where a sensitive model endpoint is reachable using a broadly shared key, the exposure is that a leaked endpoint credential might permit inference from an untrusted client; containing it requires a guardrail in the operation where identity-based access and supported network isolation reduce exposure.
Incorrect Answers
A:This restriction is associated with a blocked invoice triggers a hold on related procurement actions and addresses the exposure that a procurement approval might proceed before a finance hold is resolved; it does not specifically protect the target workflow against a leaked endpoint credential might permit inference from an untrusted client.
B:The alternate boundary mitigates a user might receive another principal's cached protected answer, which is relevant to shared result caching makes answers fast across roles; it leaves the current workflow vulnerable to a leaked endpoint credential might permit inference from an untrusted client.
C:The proposed control targets an API might trust model-generated identity claims instead of valid authorization, while the given risk scenario is a leaked endpoint credential might permit inference from an untrusted client; these exposures arise in different operational contexts.
Question 870
One enterprise rollout has 2 separately testable requirements. (1) Business use: Average accuracy hides failures on complex finance questions. Existing shortcoming: Lower-cost routing degrades a business-critical subgroup. (2) Business use: Thousands of transcripts contain recurring misunderstandings. Existing shortcoming: Analysts overlook a widespread intent failure among unique phrasings. Which 2 success metrics directly satisfy these stated cases? Choose TWO.
- Root-cause clusters converted to verified corrections
- Concessions with full recommendation-to-posting audit chain
- Quality-adjusted cost per approved business result
- Cross-role cache disclosure incidents
Explanation
Correct Answers
A:The outcome root-cause clusters converted to verified corrections measures whether the workflow where thousands of transcripts contain recurring misunderstandings is improving; the scorecard separately needs quality-adjusted cost per approved business result.
C:The outcome quality-adjusted cost per approved business result measures whether the workflow where average accuracy hides failures on complex finance questions is improving; the scorecard separately needs root-cause clusters converted to verified corrections.
Incorrect Answers
B:The alternative measures concessions with full recommendation-to-posting audit chain for the situation where service recommends a price adjustment requiring finance approval; it cannot substitute for performance outcomes in the separate named cases: average accuracy hides failures on complex finance questions; thousands of transcripts contain recurring misunderstandings.
D:The alternative measures cross-role cache disclosure incidents for the situation where shared result caching makes answers fast across roles; it cannot substitute for performance outcomes in the separate named cases: average accuracy hides failures on complex finance questions; thousands of transcripts contain recurring misunderstandings.