In Microsoft 365 Copilot, what is the primary method for creating a custom agent?

Use the Microsoft Graph API to create an agent directly in the Microsoft 365 environment.

Write a custom plugin in C# and deploy it to the Microsoft 365 admin center.

Use Copilot Studio to build and configure the agent with instructions and knowledge sources.

Create a new SharePoint site and enable the 'Agent' feature to generate an agent automatically.

A company wants to use Copilot Studio to automate a data-entry task in a legacy application that runs in a web browser. The task requires the agent to log in, navigate to a form, fill in fields, and submit. The company wants the agent to handle unexpected pop-ups during the process. What should you configure?

Configure the agent's reasoning mode to handle pop-ups.

Use MCP to connect to the legacy application.

Use a Computer Use action and configure error handling for unexpected pop-ups.

Use an API plugin to interact with the legacy application's backend.

A data science team wants to integrate a custom machine learning model they built in Microsoft Foundry into a Copilot Studio agent. The model is deployed as an endpoint in Foundry. The agent needs to call this model during conversations to make predictions. What should you do?

Create a custom connector that calls the Foundry model endpoint and add it as a plugin action in Copilot Studio.

Use Computer Use to interact with the Foundry model's web interface.

Upload the model's training data to Copilot Studio as a knowledge source.

Use Model Context Protocol (MCP) to connect the agent to the Foundry model.

A company is building a Copilot Studio agent that must automate a complex data entry task in a legacy desktop application. The task involves multiple steps and requires the agent to make decisions based on the data it sees on the screen. The company also wants the agent to be able to explain its actions to users. What should you configure?

Use Computer Use for the desktop automation and configure reasoning to enable decision-making and explanations.

Use Model Context Protocol (MCP) to connect to the desktop application.

Use voice mode to guide the user through the data entry manually.

Use a custom connector to the desktop application's API.

A company is designing an agent in Copilot Studio for customer support. The agent must handle complex queries that require multi-step reasoning and provide responses through both text and voice channels. The company wants the agent to appear more natural and human-like. What should you configure in Copilot Studio?

Connect the agent to a SharePoint knowledge base using MCP.

Enable reasoning mode and configure voice mode with a natural-sounding voice.

Use a custom model in Microsoft Foundry to improve the agent's language understanding.

Add a Computer Use action to automate responses to customer queries.

Microsoft Certified:Agentic AI Business Solutions Architect

Domain 2Objective 2

Design Extensibility of AI Solutions AB-100 Practice Questions (Page 7)

Part of the Design AI-powered business solutions domain, which accounts for 25–30% of the AB-100 exam. Microsoft does not publish an official question count, but from its 100-minute exam (~40–65 total, ~10–20 in this domain), expect 3–7 from this objective — we provide 41 practice questions to prepare you well beyond it. (estimate)

41questions here

9free pages

7concepts

25–30%of the exam

Questions 31–35

31foundation · easy
In Microsoft 365 Copilot, what is the primary method for creating a custom agent?
Select an answer first
32application · medium
A company wants to use Copilot Studio to automate a data-entry task in a legacy application that runs in a web browser. The task requires the agent to log in, navigate to a form, fill in fields, and submit. The company wants the agent to handle unexpected pop-ups during the process. What should you configure?
Select an answer first
33application · medium
A data science team wants to integrate a custom machine learning model they built in Microsoft Foundry into a Copilot Studio agent. The model is deployed as an endpoint in Foundry. The agent needs to call this model during conversations to make predictions. What should you do?
Select an answer first
34expert · hard
A company is building a Copilot Studio agent that must automate a complex data entry task in a legacy desktop application. The task involves multiple steps and requires the agent to make decisions based on the data it sees on the screen. The company also wants the agent to be able to explain its actions to users. What should you configure?
Select an answer first
35application · medium
A company is designing an agent in Copilot Studio for customer support. The agent must handle complex queries that require multi-step reasoning and provide responses through both text and voice channels. The company wants the agent to appear more natural and human-like. What should you configure in Copilot Studio?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AB-100

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Microsoft. “AB-100” is a trademark of its owner, used for identification only.

Design Extensibility of AI Solutions AB-100 Practice Questions (Page 7)

Questions 31–35

In Microsoft 365 Copilot, what is the primary method for creating a custom agent?

A data science team wants to integrate a custom machine learning model they built in Microsoft Foundry into a Copilot Studio agent. The model is deployed as an endpoint in Foundry. The agent needs to call this model during conversations to make predictions. What should you do?

Keep studying — Design AI-powered business solutions