Microsoft AB-620 Designing and Building Integrated AI Agent Solutions in Copilot Studio Exam Practice Test

Create a new agent flow artifact in the environment.

Configure the invocation trigger that allows the agent to call the agent flow.

Configure the actions that implement the required behavior.

Validate the agent flow behavior with representative test runs.

Make the agent flow available for agent usage.

An agent flow must first exist in the target environment before its trigger or processing logic can be configured. The invocation trigger is then configured so Copilot Studio can initiate the flow during a conversation. For an agent-callable flow, this is typically the When an agent calls the flow trigger. The flow should also return any required results through the corresponding response action.

After establishing the trigger, the author adds the actions that implement the required business process, such as retrieving records, calling connectors, applying conditions, or updating enterprise systems. Representative test runs should then be performed to confirm that inputs are accepted correctly, actions execute successfully, errors are handled appropriately, and expected outputs are returned to the agent.

Once the flow has been verified, it can be published and made available to the agent as a tool. Exposing an incomplete or unverified flow prematurely could cause conversation failures or incorrect tool results.

Relevant study-guide area: Plan and configure agent solutions → Create and monitor agent flows in Copilot Studio → Create an agent flow . See Create an agent flow as a tool and Call an agent flow from an agent .

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: A predefined test set with expected outcomes is the only option that satisfies consistency, grounding comparison, and repeatable regression analysis. The fixed prompts ensure that each version receives the same inputs. Verified expected responses provide the baseline against which meaning, quality, keywords, or exact output can be scored. Rerunning the same set produces comparable results over time, allowing the team to calculate whether a knowledge-base change improved or degraded performance. The manual test pane is valuable for interactive debugging and inspecting the activity map, but it does not provide the same controlled batch or historical comparison. Channel usage metrics describe adoption and operational behavior, not response correctness. Transcript keyword searches can reveal patterns after deployment but do not establish a golden answer or a controlled experiment. The evaluation method should match the content: compare-meaning for legitimate paraphrases, exact match for fixed values, keyword match for required terms, or a custom rubric for domain-specific quality. Results should be retained or exported when a longer audit history is required. Study Guide alignment: Test and manage agents > Evaluate agent performance > Create a test set; Choose an evaluation method; Review test results.

===============

Comprehensive and Detailed Explanation From [Microsoft AB-620 Study Guide]

First, create a custom solution in the development environment and add the existing agent by selecting Add existing → Agent → Agent . The solution acts as the application lifecycle management package that contains the agent and its dependent components. Required objects such as agent flows, environment variables, connection references, and custom connectors should also be included when applicable.

After the agent and its dependencies are included, export the solution from the development environment. Microsoft’s documented procedure starts from the list of unmanaged solutions and exports the solution containing the agent. During the export process, the maker can select the appropriate package type. Production deployments commonly use a managed export, but the source development solution remains unmanaged; it does not need to be converted before export.

Removing topics directly from the solution is specifically discouraged because changing or removing agent components outside the standard Copilot Studio authoring interface can cause export or import failures. Publishing the agent to a channel is also unrelated to creating a solution package. After the solution is imported into production, the imported agent must be configured, tested, and published in that environment.

Relevant study-guide area: Test and manage agents → Implement application lifecycle management for agents in Copilot Studio → Create a solution and add existing agents to a solution . See Export and import agents using solutions .

Approval stage → Wait for a response and branch according to the decision; Information-collection stage → Wait for a response and capture the submitted inputs.

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: An approval is a decision gate. The flow must pause until the reviewer responds, then branch to the appropriate approved, rejected, or other outcome path. Information collection also pauses, but its purpose is to capture supplied values and make them available to later actions in the same run. Sending a request or notification and continuing immediately would break both requirements because downstream processing could occur before the human response exists. The action should have a defined timeout and an escalation or cancellation path so a missing response does not leave the process suspended indefinitely. The flow should store the response, responder identity, timestamp, and any comments needed for audit. Sensitive information should not be requested through an inappropriate channel and should be protected in logs and variables. Testing must cover approval, rejection, corrected information, timeout, and unauthorized-responder cases. This produces a controlled human-in-the-loop design rather than a simple alert that has no effect on execution. Study Guide alignment: Plan and configure agent solutions > Create and monitor agent flows in Copilot Studio > Create a human-in-the-loop agent flow.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: The Ask with Adaptive Card node is the interactive topic element that waits for a card submission and stores the user ' s responses as variables. The card must contain input controls with unique identifiers and an Action.Submit button so the host returns the collected values to Copilot Studio. A Message node can display a card, but it is not the correct choice when the topic must collect and bind structured input in one step. Action.Execute is not consistently supported across Copilot Studio channels; Microsoft documents host-specific schema limits, and the default web chat does not support that action. Using the latest schema indiscriminately can therefore break portability. The card should target the lowest schema version required by every intended channel, validate required fields, and provide clear labels and error messages. After submission, the topic variables should be typed and checked before they are passed to tools or flows. Testing is required in each production channel because rendering and supported Adaptive Card features differ between Teams, web chat, and customer-service hosts. Study Guide alignment: Plan and configure agent solutions > Configure topics > Configure adaptive cards; Manage variables.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: Power Platform data policies are the preventive governance control for connectors, HTTP endpoints, knowledge sources, and combinations of data groups used by Copilot Studio agents. Administrators can classify or block connectors, restrict endpoint use, require authentication, and surface policy violations during authoring or publishing. This meets both enforcement and reporting requirements. Application Insights records runtime telemetry but does not prevent a maker from adding a disallowed connector. Security FAQs provide guidance only. Agent analytics can reveal behavior and usage trends but are detective, not an administrative guardrail. Data policies should be scoped deliberately at tenant or environment level and tested against representative agent configurations. When a violation occurs, Copilot Studio can block publishing and provide downloadable details identifying the policy conflict, enabling the administrator and maker to correct the design. Governance should also include least-privilege connections, environment segmentation, solution-based ALM, and review of exceptions; a broadly permissive policy followed by manual monitoring is not equivalent to enforcing the control at design and deployment time. Study Guide alignment: Plan and configure agent solutions > Plan an agent solution > Evaluate security and governance considerations.

===============

1. Define success criteria and evaluation objectives; 2. Select evaluation methods and metrics aligned to the criteria; 3. Collect interaction data and evaluation signals; 4. Review results and identify improvement opportunities.

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: Evaluation starts with a definition of success because the business objective determines what should be measured. A policy agent might prioritize grounded accuracy and citations, while a transactional agent also needs tool-selection and completion measures. Once objectives and thresholds are defined, the team selects suitable methods and metrics, such as compare-meaning, exact match, keyword match, expected tool use, or a custom quality rubric. The test is then executed to collect responses, scores, resource usage, transcripts, and activity signals. Only after evidence exists should reviewers interpret the results, isolate weak scenarios, and prioritize changes. Choosing metrics before defining success risks optimizing an irrelevant score; reviewing before collecting consistent evidence encourages anecdotal conclusions. Improvement should be iterative: change the knowledge source, instructions, tool metadata, or expected outcome one at a time and rerun the same set. This sequence produces an auditable link between business intent, measurement design, observed performance, and the remediation decision. Study Guide alignment: Test and manage agents > Evaluate agent performance > Choose an evaluation method; Review test results.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: A Copilot connector knowledge source is the appropriate integration for indexed enterprise content held in a non-Microsoft system. The connector brings authorized content into the Microsoft Graph search index and makes it available for knowledge grounding, while access controls determine which results a user may receive. Calling the third-party API from a topic on every question is a tool pattern and would require the agent to implement retrieval, ranking, and response handling itself. A Copilot Studio tool is also an executable capability, not automatically a knowledge source, so options B and C do not meet the indexed-grounding requirement. After the tenant administrator configures and synchronizes the connector, the builder adds the corresponding enterprise knowledge source to the agent, supplies clear metadata, and tests retrieval and citations. Content lifecycle, crawl status, schema mapping, and permission trimming must be monitored because a technically successful connection can still return stale, poorly ranked, or overexposed content. The design should avoid uploading a second copy that would drift from the connector-managed index. Study Guide alignment: Integrate and extend agents in Copilot Studio > Connect to enterprise knowledge sources > Connect to Copilot connectors.

===============

A. Use a Microsoft Graph connector to index the partner’s travel advisory content. Microsoft 365 Copilot connectors—formerly called Microsoft Graph connectors—index non-Microsoft enterprise content into Microsoft Graph, making it available for semantic knowledge grounding and citations. Microsoft Learn: Copilot connectors

D. Add the internal policy documents as a knowledge source. The SharePoint PDFs and policy documents should be connected as agent knowledge so generative answers can retrieve and summarize the relevant content. This is scalable and avoids manually maintaining document-specific topics. Microsoft Learn: SharePoint knowledge sources

B violates the requirement to avoid manually authored Q & A topics. C could produce ungrounded responses from unapproved public sources and would not satisfy the airline’s accuracy, security, or compliance requirements.

Requirement

Implementation method

Let the agent decide when to run a tool based on user input.

Enable generative orchestration.

Run a specific agent flow at a defined step in a topic.

Add a tool node.

Use a connector action to retrieve data during a conversation.

Add a tool node.

Comprehensive and Detailed Explanation From [Microsoft AB-620 Study Guide]

Generative orchestration is appropriate when the agent must independently determine whether a tool is relevant to the user’s request. Copilot Studio evaluates the user’s message, conversation context, and the tool’s name, description, inputs, and outputs before deciding whether to invoke it. Microsoft documentation confirms that generative orchestration allows an agent to choose tools based on their descriptions and the current user query.

When an operation must occur at a predetermined point in an authored topic, the operation should instead be inserted explicitly by using an Add a tool node. Selecting an agent flow from that node ensures that the flow executes at that exact position in the topic rather than being selected autonomously by the orchestrator.

The same node type is used when calling a connector action. From the topic canvas, the author selects Add a tool , chooses Connector , and then selects and configures the required connector operation. Adaptive Cards present interactive information, while Question nodes collect user input; neither directly invokes flows or connectors.

Relevant study-guide area: Plan and configure agent solutions → Configure topics → Add agent flows to a topic and Add tools to a topic . See Orchestrate agent behavior with generative AI , Add tools to custom agents , and Use connectors in Copilot Studio agents .

B. Configure prompt instructions that include disclaimer text. Centralized prompt instructions apply the required disclaimer and behavioral controls consistently to generative responses. This avoids maintaining duplicate text in individual topics and reduces the risk of inconsistent compliance language. Microsoft confirms that prompt modifications can control response content, format, personality, and what the agent should or should not answer. Microsoft Learn: Prompt modification

C. Use Markdown syntax within response content. The scenario explicitly requires Markdown for formatting and states that HTML is unsupported. Markdown can provide headings, bold text, lists, and other structured presentation while remaining compatible with the target channels.

Options A and E duplicate disclaimer content across topics, creating unnecessary maintenance and inconsistent responses when the disclaimer changes. Option D violates the stated platform requirement because HTML must not be used.

A. Configure prompt modifications at the system level. Centralized instructions enforce the required tone, disclaimers, response boundaries, and refusal behavior consistently across generative responses. Microsoft documents that prompt modification can define an agent’s personality, response format, permitted subjects, and behavior when an answer is unavailable. This avoids inconsistent manual changes across individual topics. Microsoft: Prompt modification

D. Configure Power Platform DLP policies. DLP policies govern which connectors, knowledge sources, HTTP endpoints, and channels the agent may use. They should be configured and validated before publication because a policy violation can prevent the agent from being published. DLP also protects against unauthorized data movement between business and non-business connectors. Microsoft: Data policies for agents

Why the others are incorrect: B creates duplicated, inconsistent disclaimers. C incorrectly treats moderation as a web-channel setting; moderation is configured for generative behavior rather than enforced only by embedding settings. E applies safeguards after publication and assumes filters must be configured separately for every channel, which would not provide the required centralized governance.

Authenticated governed access → Select service-principal authentication; Required curated vector index → Enter the Azure AI Search index name; Correct Azure AI Search service instance → Provide the Azure AI Search endpoint URL.

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: An Azure AI Search connection needs three distinct pieces of configuration. The endpoint identifies the search-service instance that hosts Fabrikam ' s indexed documents. The index name then selects the specific vector index containing the curated policies and procedures. Authentication determines how Copilot Studio is authorized to query that service. Because the scenario requires governed application access rather than a maker ' s personal identity, the service-principal option is the appropriate selection from the choices provided. Enabling citations is useful after retrieval is configured, but it does not establish the service connection or select an index. Microsoft recommends creating Azure AI Search through the supported data-source connection experience rather than manually assembling an unsupported endpoint/key combination. Administrators should grant the service principal only the permissions needed for query operations, protect its credentials, and maintain separate configuration per environment. Testing should confirm that the intended index—not another index in the same service—is queried, that unauthorized content is excluded, and that policy responses cite material that actually supports the generated guidance. Study Guide alignment: Integrate and extend agents in Copilot Studio > Connect to enterprise knowledge sources > Connect to Azure AI Search.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: Fabrikam needs both reusable prompt behavior and explicit model governance. A reusable custom prompt template provides the stable summarization instructions, expected structure, and input/output contract that can be applied wherever summarization is required. Within that prompt ' s configuration, selecting the approved model from the Microsoft Foundry model catalog controls which model generates the output. These are separate controls: the template governs the task, while the model selection satisfies the approved-model requirement. Publishing only deploys the configuration already selected; it does not automatically choose or enforce an approved model. Citations are valuable for grounded policy answers but do not determine the model used for summarization. Azure AI Search is the retrieval system for indexed enterprise policy content and is not required merely to apply a long-form summarization template. The prompt should be tested with representative transcripts for completeness, classification accuracy, output schema, safety, and latency. Where Fabrikam already uses a connected Foundry agent for this specialist function, the prompt and model settings should remain owned by that component rather than being duplicated in the primary agent. Study Guide alignment: Integrate and extend agents in Copilot Studio > Integrate agents with Azure > Configure custom prompts to use the Foundry model catalog.

===============

Prevent most builders from modifying shared assets → Share the component collection while controlling edit access; Reuse components without copying → Create a component collection and move the shared components into it; Limit use to the intended primary agent → Restrict the component collection to the primary agent.

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: A component collection is the correct reusable unit for the escalation topics, MCP definitions, and safety disclaimer. Moving those assets into one collection creates a maintained source that multiple agents can consume without copying and later diverging. Sharing the collection with teammates while controlling edit access separates use permission from modification permission, allowing agent authors to reference the assets while reserving changes for platform engineering. If a sensitive or tightly coupled collection must be usable only by one designated primary agent, the collection can be restricted accordingly. Exporting the primary agent and copying its components would recreate the duplication problem that Fabrikam is trying to eliminate, while restricting every collection to the primary agent would prevent the required cross-agent reuse. The collection should be versioned and tested after changes because a shared update can affect several consumers. Ownership, review, and release responsibilities should also be documented so that reuse does not become an uncontrolled dependency. This design provides consistency, centralized governance, and a smaller maintenance surface across the three agents. Study Guide alignment: Plan and configure agent solutions > Plan an agent solution > Plan reusable agent components.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: The Service Desk Agent must be both published and available for connections from other agents. Publishing establishes a runtime version that the primary agent can call; an unpublished draft is not a valid connected-agent target. On the Service Desk Agent, the setting that allows other agents to connect and use it must also be enabled. Microsoft additionally requires the connected agent to be in the same environment and the maker to own it or have it shared, but those conditions are not offered as choices. Replacing the agent with an MCP server would discard the existing agent ' s orchestration and ticket-routing logic. Adding it as a knowledge source would expose content for retrieval, not delegate an IT request to another agent. Publishing Operations Concierge alone does not satisfy the prerequisites of the target agent. After connection, the local description in Operations Concierge should clearly define when IT-specific work is delegated, because descriptions influence routing. Any later changes to the Service Desk Agent must be published before the main agent can use the updated version. Study Guide alignment: Integrate and extend agents in Copilot Studio > Configure multi-agent collaboration from Copilot Studio > Integrate an existing agent in Copilot Studio.

===============

PartsOps per-user authentication → OAuth 2.0; WarrantyRules shared team credential → API key; Tool-selection context → Configure a server description that explains the MCP server ' s role.

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: The authentication choices must follow the identity requirements of each server. PartsOps performs attributable actions, including parts-request creation, so OAuth 2.0 is the appropriate mechanism for individual users to authorize access without sharing their credentials with the agent. That preserves user context and supports accountability for each transaction. WarrantyRules is explicitly read-only and uses a key shared by the agent team; it should therefore be configured with API-key authentication using the required header or query parameter. Applying the same authentication method to both servers would contradict the stated server designs. Authentication alone does not tell the orchestrator when a server is relevant. A concise, discriminating server description should identify PartsOps as the source for availability checks and parts-request creation and WarrantyRules as the source for warranty validation. Copilot Studio uses the names and descriptions of agents, topics, tools, and knowledge sources during generative orchestration. Accurate metadata reduces accidental calls, prevents the wrong server from handling a request, and makes the multi-agent plan more predictable during testing. Study Guide alignment: Integrate and extend agents in Copilot Studio > Add tools to agents > Configure MCP tools.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: A Fabric Data Agent is the capability designed for questions that must be answered from governed analytical data. Fabrikam ' s backlog, SLA-risk, parts-shortage, and dispatch-delay measures already reside behind a Fabric semantic model, so delegating these requests to the data agent preserves the business definitions, relationships, and security applied in Fabric. The result can be returned in a structured form suitable for weekly operational review rather than being estimated from general model knowledge. A Foundry agent is already assigned to summarization and classification of long-form text; it is not the authoritative source for operational KPIs. An MCP server is appropriate when the user must perform a transaction, such as creating a parts request, but the question asks for performance information rather than an update. Azure AI Search supports document retrieval for policies and procedures, not calculation over the Fabric semantic model. The primary agent ' s description for the Fabric specialist should explicitly mention the supported metrics so generative orchestration consistently delegates those requests instead of attempting an unsupported generative answer. Study Guide alignment: Integrate and extend agents in Copilot Studio > Configure multi-agent collaboration from Copilot Studio > Integrate a Fabric data agent.

===============

Comprehensive and Detailed Explanation From Microsoft AB-620 Study Guide: The partner capability is described as an agent available only through a standardized agent-to-agent endpoint. That is the defining use case for an Agent2Agent connection. In Copilot Studio, the primary agent is configured with the external agent ' s A2A message endpoint and the applicable authentication details, after which orchestration can delegate shipment-tracking requests to it. An MCP server exposes tools or resources rather than an independently orchestrated partner agent, so option C changes the integration contract. An indexed knowledge source is also unsuitable because shipment status is dynamic and the requirement is delegation, not retrieval from a copied corpus. Generative responses in the primary agent would have no authoritative live shipment source and would violate the defined architecture. The A2A description should identify shipment tracking precisely so that it does not compete with unrelated agents or tools. Because external-agent support can be preview functionality, Fabrikam should also validate regional availability, authentication, error behavior, latency, and governance before production use. Study Guide alignment: Integrate and extend agents in Copilot Studio > Configure multi-agent collaboration from Copilot Studio > Create a multi-agent solution by using A2A protocol.

===============