View Full Microsoft AB-100 Exam Dumps and Practice Test Dumps.

 

Question 21

What is a key benefit of using a small language model for a focused business task?

1. It always provides better results than every larger model
2. It eliminates the need for business data
3. It can provide efficient performance for a specific use case
4. It removes the need for testing

Correct Answer: 3

Explanation

Small language models can be useful when a business task has a focused scope and does not require the broad capabilities of a larger model. They may provide advantages in areas such as cost, latency, resource usage, and deployment flexibility. The architect should evaluate the task requirements, expected accuracy, model capabilities, and operational constraints before selecting a model. A smaller model should not be selected simply because it is cheaper; it must still meet the required business and quality criteria.

Question 22

Which approach helps an AI solution use company-specific information?

1. Grounding
2. Screen sharing
3. File compression
4. Network routing

Correct Answer: 1

Explanation

Grounding connects an AI solution with relevant information that can improve its responses for a specific business scenario. Sources may include approved business documents, application data, or other organizational knowledge. The architect should assess the quality, relevance, freshness, and availability of the information before using it. Access permissions are also important because the AI solution should not expose information to users who are not authorized to view it. Effective grounding helps connect general AI capabilities with the organization’s actual business context.

Question 23

Which agent is designed to perform actions with a high degree of independence?

1. Prompt agent
2. Task agent
3. Autonomous agent
4. FAQ agent

Correct Answer: 3

Explanation

An autonomous agent is designed to operate with a greater degree of independence when completing objectives. It may determine steps, use available tools, and take actions without requiring confirmation for every individual operation. Because greater autonomy can introduce additional risks, the architect must define appropriate boundaries, permissions, safeguards, monitoring, and escalation mechanisms. Autonomous behavior should be justified by the business requirement. For simpler and more controlled scenarios, a task or prompt-based agent may be more appropriate.

Question 24

A company wants an agent to answer questions using internal documents. What should the architect define first?

1. Monitor size
2. Knowledge sources
3. Office locations
4. Printer settings

Correct Answer: 2

Explanation

Knowledge sources determine which information an agent can use when responding to business questions. The architect should identify authoritative sources and evaluate their accuracy, relevance, freshness, cleanliness, and availability. The design should also address access permissions and how information will be retrieved. Selecting the right knowledge sources helps ensure that responses are based on appropriate organizational information. The architect should avoid using unverified or outdated content because poor grounding data can negatively affect the reliability and usefulness of the agent.

Question 25

Which technology can provide standardized access between an AI application and external tools or data sources?

1. Model Context Protocol
2. SMTP
3. DHCP
4. FTP

Correct Answer: 1

Explanation

Model Context Protocol, or MCP, is an open protocol that can help AI applications connect with external tools and context in a standardized way. This can support agent extensibility by allowing an AI solution to interact with supported capabilities without creating a completely separate integration pattern for every tool. Architects must still evaluate authentication, authorization, data exposure, reliability, and governance. MCP provides an interoperability mechanism, but it does not replace the security and operational controls required for an enterprise AI solution.

Question 26

What should be included when calculating the ROI of an AI solution?

1. Only the purchase price
2. Only employee salaries
3. Benefits and total costs
4. Only model accuracy

Correct Answer: 3

Explanation

ROI analysis should consider both the expected business benefits and the total costs associated with an AI solution. Costs can include development, licensing, infrastructure, integration, support, maintenance, and operational expenses. Benefits may include productivity improvements, reduced processing costs, faster service, improved customer experiences, or increased revenue. The architect should define measurable criteria so the organization can compare expected outcomes with the investment required. Considering total cost of ownership helps create a more realistic view of the solution’s financial impact.

Question 27

Which Copilot Studio feature can help an agent perform a specific business operation?

1. Actions
2. Wallpapers
3. Fonts
4. Screensavers

Correct Answer: 1

Explanation

Actions allow an agent to perform operations or interact with connected business capabilities. They can help an agent move beyond simply generating responses and enable it to participate in business processes. When designing actions, the architect should consider required inputs, outputs, permissions, error handling, and security boundaries. Actions should only provide the capabilities necessary for the intended business scenario. Limiting permissions and clearly defining action behavior helps reduce the risk of unintended operations while keeping the agent aligned with business requirements.

Question 28

A business wants to use AI in a Power Apps canvas app. What should the architect consider?

1. Only the app icon
2. The AI component and business workflow
3. Only screen colors
4. Only user passwords

Correct Answer: 2

Explanation

When AI is incorporated into a Power Apps canvas app, the architect should consider how the AI capability fits into the existing business workflow. This includes identifying the task AI will perform, the information it requires, the expected output, and how users will interact with the result. Security, data access, performance, and error handling should also be considered. AI should solve a meaningful business problem rather than being added simply because it is available. The overall application design should remain understandable and maintainable.

Question 29

Which approach can help select the most suitable AI model for different requests?

1. Model routing
2. Manual file copying
3. Network printing
4. Password rotation

Correct Answer: 1

Explanation

Model routing allows an AI solution to select an appropriate model based on characteristics of the request. Factors can include task complexity, required capabilities, response time, cost, and quality requirements. A routing strategy can help organizations use more capable models when necessary while directing simpler requests to less expensive or faster models. The architect should define routing rules and fallback behavior and monitor results to verify that the selected models meet the expected quality and performance requirements.

Question 30

Which approach is most appropriate for a predictable, narrowly defined business task?

1. Unlimited autonomy
2. Task agent
3. Uncontrolled multi-agent system
4. Manual processing only

Correct Answer: 2

Explanation

A task agent is well suited to a business process with a defined objective and predictable sequence of activities. The architect can establish clear inputs, actions, outputs, permissions, and success criteria. This focused design can make the solution easier to test, monitor, and govern than an unnecessarily autonomous architecture. The architect should determine whether the task requires AI at all and select the simplest approach that meets the requirement. Greater autonomy should be introduced only when the business scenario genuinely requires it.

Question 31

What is prompt engineering primarily concerned with?

1. Designing effective instructions for AI models
2. Managing physical servers
3. Configuring printers
4. Creating network cables

Correct Answer: 1

Explanation

Prompt engineering involves designing instructions that help an AI model produce useful and consistent results. Effective prompts can define the task, provide relevant context, specify constraints, and describe the desired output format. Architects should encourage testing and refinement because prompt behavior can vary depending on the model and scenario. Prompt guidelines can also improve consistency across an organization. When prompts are used in business solutions, they should be reviewed alongside security, responsible AI, data access, and other architectural requirements.

Question 32

An agent needs to respond differently when it cannot understand a user’s request. What should the architect design?

1. A fallback path
2. A larger monitor
3. A new printer
4. A database backup

Correct Answer: 1

Explanation

A fallback path provides an alternative response or process when an agent cannot confidently understand or fulfill a request. In Copilot Studio, fallback behavior can help direct users toward clarification, supported topics, another process, or human assistance. The architect should define what happens when the agent cannot determine the user’s intent or when an action fails. Good fallback design improves user experience and reduces unpredictable behavior. It also provides an opportunity to clearly communicate limitations rather than allowing the agent to produce unsupported responses.

Question 33

Which Microsoft solution is commonly used to build low-code agents?

1. Microsoft Copilot Studio
2. Microsoft Excel
3. Microsoft Paint
4. Windows Notepad

Correct Answer: 1

Explanation

Microsoft Copilot Studio provides a low-code environment for creating and customizing agents. It supports capabilities such as topics, knowledge, actions, prompts, and agent flows. This makes it useful for organizations that want to build business-focused agents without implementing every component through traditional software development. The architect should still evaluate the required capabilities, integrations, security, governance, and scalability. If the scenario requires more advanced customization or specialized model development, other Microsoft AI services may be considered.

Question 34

What should an architect evaluate before selecting a grounding source?

1. Its relevance and data quality
2. Its file name length
3. Its screen resolution
4. Its office location

Correct Answer: 1

Explanation

Grounding sources should be evaluated for relevance and quality before being incorporated into an AI solution. Important factors include accuracy, freshness, cleanliness, availability, and whether the source is authoritative for the business scenario. The architect should also consider how access permissions are enforced and whether the data can be retrieved effectively. Using irrelevant or unreliable information can reduce response quality and increase the risk of incorrect answers. Data preparation and governance are therefore important parts of an effective grounding strategy.

Question 35

Which protocol is designed for communication between AI agents?

1. HTTP
2. Agent2Agent
3. SMTP
4. DNS

Correct Answer: 2

Explanation

Agent2Agent, or A2A, is an open protocol designed to support communication and collaboration between AI agents. It can be useful in multi-agent architectures where specialized agents need to interact to complete a broader business objective. The architect should define responsibilities, communication patterns, security requirements, and error-handling behavior for participating agents. Using an appropriate interoperability standard can reduce the need for tightly coupled custom integrations. However, protocol support does not eliminate the need for identity, authorization, monitoring, and governance controls.

Question 36

Why should an AI solution have defined rules and constraints?

1. To increase screen resolution
2. To control expected agent behavior
3. To reduce storage space
4. To replace all testing

Correct Answer: 2

Explanation

Rules and constraints help establish boundaries for how an AI solution should behave. They can specify permitted actions, acceptable responses, restricted data, escalation requirements, and conditions that require user confirmation. Clearly defined boundaries are particularly important for agents that can interact with business systems or take actions on behalf of users. Architects should combine behavioral constraints with identity, authorization, monitoring, testing, and governance controls. This helps keep the AI solution aligned with business requirements and reduces the risk of unintended behavior.

Question 37

Which factor is important when deciding between a custom AI model and an existing model?

1. Business requirements
2. Keyboard brand
3. Office furniture
4. Monitor color

Correct Answer: 1

Explanation

The decision to create a custom AI model should be based on business and technical requirements. The architect should determine whether existing models can provide the required quality, domain knowledge, performance, cost, and capabilities. Customization may be justified when specialized requirements cannot be adequately addressed by available models. However, custom models can introduce additional development, testing, maintenance, and governance responsibilities. A build-versus-buy analysis helps organizations determine whether customization provides enough value to justify its additional complexity and ongoing operational costs.

Question 38

Which activity helps determine whether an AI agent meets business requirements before production?

1. Agent testing
2. Logo redesign
3. Printer replacement
4. Office relocation

Correct Answer: 1

Explanation

Agent testing verifies whether an AI solution behaves as expected under defined scenarios. Tests can evaluate response quality, task completion, accuracy, safety, performance, and failure handling. The architect should establish measurable validation criteria before testing begins. Testing should include normal requests as well as edge cases and potentially harmful or unexpected inputs. For solutions involving multiple agents or business applications, end-to-end testing is also important. Effective testing reduces deployment risk and helps identify areas requiring prompt, model, workflow, or architecture changes.

Question 39

What is a key purpose of telemetry in an AI solution?

1. Monitoring behavior and performance
2. Changing application colors
3. Creating user passwords
4. Managing office equipment

Correct Answer: 1

Explanation

Telemetry provides information about how an AI solution behaves during operation. It can help architects and administrators analyze performance, errors, usage patterns, response times, and other relevant indicators. This information supports troubleshooting, tuning, and continuous improvement. For agentic solutions, telemetry can also help identify unexpected behavior or inefficient workflows. Monitoring should be designed with appropriate privacy and security controls. The collected information should support meaningful operational decisions rather than simply generating large volumes of data without a defined purpose.

Question 40

A company wants to move an AI solution from development to production in a controlled manner. What should be used?

1. Random manual changes
2. An ALM strategy
3. Uncontrolled user access
4. No testing

Correct Answer: 2

Explanation

An application lifecycle management strategy provides a controlled approach for developing, testing, deploying, and maintaining an AI solution. For agentic solutions, ALM can cover agents, connectors, actions, prompts, data, configurations, and other related components. The architect should define environments, version control, testing, deployment processes, approvals, rollback procedures, and governance. Separating development and production helps reduce the risk of untested changes affecting users. A structured ALM process also supports repeatable deployments and long-term maintenance of the AI solution.