Microsoft AB-730 Online Practice Questions and Exam Preparation

Microsoft AB-730 Online Questions & Answers

• Question 41:

  You ask Microsoft 365 Copilot to create a report based on information from the web.

  You verify the response and discover that some information is fictional.

  What is this an example of?

  A. deepfake
  B. fabrication
  C. overreliance
  D. prompt injection
  E. bias
  B. fabrication

  ---

  Explanation

  Fabrication is when an AI system produces information that appears plausible but is not grounded in the available evidence. In this scenario, the user checks the response and finds that some web-based information is fictional. That is different from overreliance, which is the user's failure to verify, and different from prompt injection, which involves malicious instructions or context manipulation. The practical rule is to verify important claims because AI-generated content can include unsupported or invented details.

• Question 42:

  You are comparing the difference between Microsoft 365 Copilot and Microsoft 365 Copilot Chat.

  What is available in both versions?

  A. the Researcher agent
  B. Copilot Pages
  C. Copilot Notebooks
  B. Copilot Pages

  ---

  Explanation

  Copilot Pages is available in Microsoft 365 Copilot Chat experiences and provides a persistent canvas for turning Copilot responses into editable, shareable work. Researcher and Notebooks are advanced Microsoft 365 Copilot capabilities that depend on licensing and are not available in the same way in every Copilot Chat version. The question asks for a feature available in both versions, and Pages is the shared collaboration surface. The exam takeaway is to separate broadly available Chat collaboration features from advanced licensed agents and notebook capabilities.

• Question 43:

  HOTSPOT

  Select the answer that correctly completes the sentence.

  

  

  ---

  The selected completion says you cannot modify the settings of the Researcher agent. Researcher is a Microsoft-provided advanced agent, not a custom agent that the user owns and configures. Users can run Researcher for research tasks, but they do not edit its underlying settings, instructions, or suggested prompts as if it were their own custom agent. The distinction is between built-in Microsoft agents and agents you create or administer.

• Question 44:

  You are a medical researcher.

  You need to ensure that Microsoft 365 Copilot responds to you in a professional manner that is appropriate to your field.

  What is the best approach to achieve the goal?

  More than one answer choice may achieve the goal. Select the BEST answer.

  A. In the conversation, attach an example to your prompts.
  B. From Settings, configure Custom instructions.
  C. In the conversation, create a sample prompt and save it to the Copilot Prompt Gallery.
  D. From Settings, configure Copilot Memory.
  B. From Settings, configure Custom instructions.

  ---

  Explanation

  Custom instructions let the user define how Copilot should respond across conversations, including preferences for tone, role, style, and professional context. A medical researcher who needs responses in an appropriate professional manner should configure those instructions rather than attaching examples every time. Copilot Memory stores remembered details, but custom instructions are the direct control for response style and expectations. The exam takeaway is to use custom instructions for persistent response preferences.

• Question 45:

  HOTSPOT

  Select the answer that correctly completes the sentence.

  

  

  ---

  The selected completion is generate a recap of what was discussed during the meeting. After a Teams meeting, Copilot's after-meeting usefulness depends heavily on whether the meeting was transcribed. A transcript enables recap, summaries, and follow-up questions based on what was discussed. Scheduling follow-up meetings or identifying attachments can be separate tasks, but the selected after-meeting Copilot function is recap generation from the transcript.

• Question 46:

  A business analyst needs a structured, source-cited report that combines information from the web with work files, emails, meetings, and chats that the analyst is allowed to access.

  The task requires deeper multistep research, not a short chat answer.

  Which Microsoft 365 Copilot capability should the analyst use?

  A. Researcher
  B. Copilot Pages
  C. Themes by Copilot in Outlook
  D. Meeting Recap
  A. Researcher

  ---

  Explanation

  Researcher is designed for deeper, source-cited research that can combine web information with work context the user is permitted to access. The scenario requires a structured report and multistep research rather than a short answer or a meeting recap. Copilot Pages can hold and collaborate on generated content, but it is not the specialized research agent. Use Researcher when the task is to investigate, synthesize, and cite sources across multiple information types.

• Question 47:

  You are using Microsoft 365 Copilot to enhance your productivity in Microsoft Word.

  Which two tasks can you achieve by using Copilot in Word?

  Each correct answer presents a complete the solution.

  NOTE: Each correct selection is worth one point.

  A. Insert a table of contents based on the headings in your document.
  B. Customize the page margins to match your company's document standards.
  C. Generate a summary of the key points in your document.
  D. Insert a custom watermark that has specific text and formatting.
  A. Insert a table of contents based on the headings in your document.
  C. Generate a summary of the key points in your document.

  ---

  Explanation

  Copilot in Word can summarize a document and help draft or add text based on document context, referenced files, or a user prompt. The supported capability that is clearly represented in the answer choices is generating a summary of the key points in the document. Table of contents insertion, page margin configuration, and custom watermark formatting are conventional Word layout features and are not reliable examples of Copilot generating or reasoning over document content. Because the question asks for two selections but only one listed task is clearly supported by the cited Copilot in Word behavior, the safest correction is to keep only the document-summary task.

• Question 48:

  HOTSPOT

  You are writing Microsoft 365 Copilot prompts.

  For each prompt need shown in the exhibit, select the prompt component that best matches the requirement.

  NOTE: Each correct selection is worth one point.

  

  

  ---

  The selected mappings are Goal, Source, Context, and Expectations. Stating what Copilot should produce is the goal, while identifying files, emails, meetings, or web data tells Copilot the source to use. Explaining the audience, business situation, or background provides context. Defining format, tone, length, or success criteria sets expectations for the response.

• Question 49:

  You receive several images from a colleague.

  You suspect that the images were generated by using Microsoft 365 Copilot.

  What can you use to verify whether the images were AI-generated?

  A. file names
  B. watermarks
  C. content credentials
  D. file descriptions
  C. content credentials

  ---

  Explanation

  Content credentials are designed to provide provenance information for digital content, including whether an image was generated or edited with AI. Microsoft states that AI-generated images created in Microsoft 365 Copilot can include content credentials based on the C2PA provenance standard. File names and descriptions can be changed easily, and a visible watermark is not the reliable mechanism being tested here. For verification, students should look for content credentials rather than relying on cosmetic or user-editable metadata.

• Question 50:

  HOTSPOT

  Select the answer that correctly completes the sentence.

  

  

  ---

  The selected completion is the Configure tab. In the agent creation flow, the Configure tab is where the builder defines key agent configuration such as instructions, knowledge, and capabilities. Notebooks and pages are content or workspace features, and the Describe tab is not the selected build surface in the exhibit. The exam point is that generative AI-assisted agent building is controlled from the agent configuration experience.