Agentic AI Business Solutions Architect · 46% of the exam

Deploy AI-powered business solutions: free practice questions

5 sample questions from our 21-question bank for this domain — answers and explanations included. These are the same scenario-based style as the real Microsoft exam.

1. Contoso Financial is building a validation framework for a custom Azure Machine Learning regression model that predicts credit-limit recommendations surfaced through a Dynamics 365 Sales Copilot panel. The model will be retrained monthly on new transaction data. The Responsible AI review board requires that validation criteria be formally defined before the model is promoted to production after each retraining cycle. Which TWO criteria should the architect include as mandatory gates in the model validation process? (Select TWO.)

  • A. The retrained model must achieve equal-or-better performance on a held-out fairness dataset segmented by protected demographic groups compared with the currently deployed model version.✓ Correct
  • B. The model's Docker container image must pass a vulnerability scan with zero critical CVEs before the model endpoint is updated in Azure Machine Learning.
  • C. A human-in-the-loop review must be completed for every individual prediction the model makes before the result is displayed in Dynamics 365 Sales.
  • D. The retrained model must meet or exceed baseline regression metrics (e.g., RMSE and MAE) on the designated test dataset before promotion.✓ Correct
  • E. The model training run must complete within a fixed compute-cost budget defined in the Azure Machine Learning compute cluster configuration.
  • F. All training data must be re-labelled by a human reviewer after each monthly ingestion to eliminate potential label drift before retraining begins.
Explanation

Fairness evaluation against protected demographic groups (A) is a mandatory Responsible AI gate that ensures the model does not produce biased credit recommendations — a regulatory requirement in financial services. Baseline regression metric thresholds (D) such as RMSE and MAE are standard quantitative quality gates that confirm the retrained model has not regressed in predictive accuracy before promotion. Together these address both performance and fairness. — Container vulnerability scanning (B) is an important DevSecOps practice but is an infrastructure gate, not a model validation criterion; it does not validate model quality or fairness. — Human-in-the-loop review for every prediction (C) is operationally infeasible at scale and is not a validation criterion for the model itself; human oversight is better applied at high-risk decision thresholds. — Compute cost budget compliance (E) is a cost-governance concern, not a model quality or responsible AI validation gate. — Requiring full human re-labelling of all data monthly (F) is disproportionate, not a standard validation criterion, and would make the monthly retraining cycle impractical.

2. A government agency is deploying a Microsoft Copilot Studio agent that will access citizen records stored in Dataverse. The agency's compliance team requires that all changes to the grounding data used by the agent — including record creation, modification, and deletion — be traceable to individual user identities. Which capability should the architect configure to satisfy this requirement?

  • A. Dataverse auditing with entity-level and field-level audit logging enabled for relevant tables✓ Correct
  • B. Power Platform Data Loss Prevention (DLP) policies scoped to the agent's environment
  • C. Azure Purview data catalog scans scheduled weekly on the Dataverse connector
  • D. Managed identity assignment on the Copilot Studio agent's Power Platform environment
Explanation

Dataverse auditing captures who created, modified, or deleted records at both the entity and field levels, providing the user-attributed audit trail required by the compliance team. DLP policies control which connectors and data flows are permitted but do not log individual data record changes to specific users. Azure Purview catalog scans classify and discover data assets for governance metadata purposes — they do not produce change-level audit logs. Managed identity assignment governs how the agent authenticates to downstream services but does not itself create an audit trail of data changes.

3. A Power Platform architect is designing the ALM process for custom connectors used by a Copilot Studio agent solution that will be deployed across 15 customer tenants. Which TWO practices should be mandated in the ALM process to ensure consistency and security across tenants?

  • A. Store custom connector definitions as solution components in source control and deploy them as part of a managed solution, so the connector schema is locked in each tenant's production environment✓ Correct
  • B. Include automated OpenAPI schema validation in the CI/CD pipeline to detect breaking changes to connector endpoints before deployment✓ Correct
  • C. Require tenant administrators to manually recreate connector configurations in each environment using the Power Platform admin center UI
  • D. Grant each tenant's global administrator the ability to modify the managed connector definition directly in their production environment to accommodate local customizations
  • E. Use a single shared API key embedded in the connector definition for all 15 tenants to simplify credential management
Explanation

Storing connector definitions as managed solution components in source control ensures version control, auditability, and environment lock-down — tenants cannot accidentally modify a production-managed connector. Automated OpenAPI schema validation in the CI/CD pipeline catches breaking changes (e.g., renamed parameters, removed endpoints) before they affect any tenant. Manual recreation across 15 tenants is error-prone, inconsistent, and unscalable. Allowing tenant administrators to modify managed connector definitions defeats the purpose of managed solutions and breaks the central governance model. Embedding a shared API key in the connector definition is a severe security anti-pattern: a single compromised key affects all 15 tenants and the key is visible to anyone who exports the solution.

4. AdventureWorks has deployed a Copilot Studio agent that integrates with Dynamics 365 Field Service. Over the past month, service managers have submitted backlog feedback indicating the agent frequently misclassifies service request urgency. The architect wants to use AI-based tools to analyze the issue systematically. Which sequence of actions is MOST appropriate?

  • A. Export agent conversation transcripts, use Azure AI Language or a similar tool to identify patterns in misclassified utterances, correlate findings with the agent's topic-level analytics, and then retune the relevant topics or update the classification prompt✓ Correct
  • B. Immediately republish the agent with a modified system prompt that includes all historical misclassified examples as few-shot samples, then monitor for 24 hours
  • C. Disable the urgency classification feature entirely and route all service requests to a human dispatcher until a root cause is determined
  • D. Increase the agent's session timeout value and add a fallback topic that captures unrecognized intents, then redeploy without further analysis
Explanation

A systematic analytical approach — exporting transcripts, applying NLP tools to identify misclassification patterns, correlating with built-in analytics, then applying targeted remediations — is the recommended methodology for AI-based issue analysis. It ensures the fix addresses root causes rather than symptoms. Immediately republishing with all misclassified examples as few-shot prompts risks overfitting to historical errors and bypasses proper testing. Disabling the feature entirely is operationally disruptive and unnecessary before root cause analysis is complete. Increasing session timeout and adding a fallback topic addresses unrecognized intent routing, not urgency misclassification — it is a mismatch between the observed symptom and the proposed fix.

5. A manufacturing company has deployed a Dynamics 365 AI feature (predictive maintenance) using a custom AI model fine-tuned in Azure AI Foundry. The model must be retrained quarterly with new sensor data. Which ALM consideration is MOST critical when designing the pipeline for this model?

  • A. Versioning every trained model artifact with associated training dataset snapshots and evaluation metrics so that any production version can be audited, reproduced, or rolled back✓ Correct
  • B. Scheduling model retraining to run on the same compute cluster used for production inference to minimize infrastructure costs
  • C. Storing the trained model weights exclusively in a Dataverse table to keep all AI assets within the Power Platform boundary
  • D. Replacing the previous model version immediately upon completion of retraining, without a staging validation phase, to minimize the gap between training and production
Explanation

Full model versioning — linking each artifact to its training dataset snapshot and evaluation metrics — is the cornerstone of MLOps ALM. It enables auditability (regulators or internal reviews can inspect what data produced which model), reproducibility, and safe rollback if a new version underperforms. Using the same compute cluster for training and inference risks resource contention that could degrade production predictions. Storing model weights in Dataverse is inappropriate; model artifacts belong in blob-backed model registries (Azure ML or AI Foundry) designed for large binary assets. Deploying without a staging validation phase bypasses quality gates, risking production degradation with every quarterly cycle.

16 more questions in this domain

Practice the full bank with instant grading, flashcards, and a timed mock exam.

Start practicing free