Free Microsoft AB-730 Practice Exam: Microsoft Certified: AI Business Professional

This free full-length Microsoft AB-730 practice exam includes 50 original IT Mastery questions across the exam domains.

These are original IT Mastery practice questions. They are not official Microsoft questions, copied live-exam content, or exam dumps. Use them for self-assessment, scope review, and deciding what to drill next.

Count note: this page uses the full-length practice count maintained in the Mastery exam catalog. Some certification vendors publish total questions, scored questions, duration, or unscored/pretest-item rules differently; always confirm exam-day rules with the sponsor.

Exam snapshot

• Exam route: Microsoft AB-730
• Practice-set question count: 50
• Time limit: 45 minutes
• Practice style: mixed-domain diagnostic run with answer explanations

Full-length exam mix

Use this as one diagnostic run. IT Mastery gives you timed mocks, topic drills, analytics, code-reading practice where relevant, and interactive practice.

Practice questions

Questions 1-25

Question 1

Topic: Draft and Analyze Business Content by Using AI

A sales operations analyst used Analyst in Excel to identify stalled Q3 pipeline opportunities. The goal is to turn the analysis into a clear Outlook email for regional sales managers.

Draft prompt: “Make this useful for the team.”

Which prompt improvement would best help Microsoft 365 Copilot create the needed business output?

Options:

• A. Using the Q3 pipeline analysis in this workbook, draft an Outlook email to regional sales managers with the top three stalled-opportunity patterns, recommended next steps, owners, and a request for source-row verification before sending.

• B. Write something persuasive about improving sales performance.

• C. Copy all customer names, deal notes, and opportunity details into an email for the whole company.

• D. Create a generic presentation about pipeline management without using the workbook.

Best answer: A

Explanation: A strong Copilot prompt for moving insights between Microsoft 365 apps should name the source analysis, the destination output, the audience, and the expected structure. Here, the user needs to transform Excel analysis into an Outlook email, so the improved prompt should keep Copilot grounded in the workbook, specify the email audience, and request actionable items such as patterns, next steps, and owners. Adding a verification step is also appropriate because pipeline insights may affect business decisions and should be checked against the source rows before sending. The key is to convert analysis into a usable communication, not ask for a vague summary or expose unnecessary data.

• Vague sales request lacks the source, audience, format, and specific action needed for a useful email.
• Overexposing details risks sharing more customer and deal information than the audience needs.
• Generic presentation uses the wrong output and ignores the workbook analysis that should drive the message.

Question 2

Topic: Understand Generative AI Fundamentals

A product marketing manager uses Microsoft 365 Copilot to draft a competitor update for executives. The update must be based on an approved internal strategy document and a recent public article, and the manager must reduce the risk of unsupported or fabricated claims. Which action best satisfies the requirement?

Options:

• A. Share the first Copilot draft if it sounds executive-ready.

• B. Reference the approved sources, request citations, and review them before sharing.

• C. Ask Copilot to create the update from memory without source references.

• D. Use only the public article so the update avoids internal context.

Best answer: B

Explanation: A responsible AI practice for business productivity is to verify AI-generated content before relying on it. For an executive competitor update, the manager should provide or reference the relevant approved sources, ask Copilot for citations or source references, and compare key claims against those sources before sharing. This reduces the risk of fabrications while keeping the output grounded in the intended business context. A polished draft is not enough; accuracy still requires source checks and human review.

• No source references fails because Copilot output can sound confident while still containing unsupported claims.
• Public article only omits the approved internal strategy document that the update must use.
• First draft sharing relies on tone instead of verifying whether the claims are source-supported.

Question 3

Topic: Manage Prompts and Conversations by Using AI

A marketing manager used Microsoft 365 Copilot to compare campaign options, choose a target segment, and draft a launch email. She wants to preserve the useful analysis and decisions for the team before moving to other work.

Draft prompt: “Remind me what we decided.”

Which improvement best supports the goal?

Options:

• A. Search every company file for launch ideas and make a long report.

• B. Delete the chat after copying the final email draft.

• C. Summarize this conversation into decisions, reusable drafts, analysis, and next steps; add it to the launch notebook.

• D. Start a new chat and create a fresh campaign plan from memory.

Best answer: C

Explanation: Conversation history is useful when a Copilot chat contains decisions, draft content, reasoning, or analysis that should not be lost. In this scenario, the manager already has valuable context in the current conversation: campaign comparisons, a chosen segment, and a draft email. A better prompt should ask Copilot to structure that existing history into useful sections and preserve it in a notebook for later reference and team follow-up.

The key idea is to continue from the current conversation when the prior context matters, rather than restarting or keeping only one output.

• Starting over loses the useful context and decisions already captured in the current conversation.
• Searching broadly is overbroad and does not focus on preserving the completed discussion.
• Copying only the draft keeps one artifact but loses the decision rationale and analysis.

Question 4

Topic: Understand Generative AI Fundamentals

A business analyst asks Microsoft 365 Copilot to create a staffing-risk summary from project files. Copilot returns a short answer and says it cannot use one referenced compensation workbook. What is the best improvement to the next prompt?

Options:

• A. Ask Copilot to search the web for similar compensation data.

• B. Ask Copilot to ignore permission limits and infer the missing compensation details.

• C. Ask Copilot to summarize accessible files, cite sources, and list inaccessible references.

• D. Paste the full compensation workbook into the chat for complete context.

Best answer: C

Explanation: A restricted or incomplete Copilot response can be caused by data protection controls, such as file permissions or sensitivity protections. The best prompt improvement is to work within accessible content, ask for citations, and have Copilot identify which referenced sources could not be used. That helps the analyst verify whether the gap is due to missing context, lack of access, or protected information. If the workbook is needed, the appropriate next step is to obtain approved access or an approved excerpt through normal business processes. The key is not to bypass protections or ask Copilot to guess sensitive details.

• Ignoring permissions is unsafe because Copilot should not bypass organizational access controls or fabricate missing information.
• Pasting sensitive data may expose protected compensation information without confirming that sharing is appropriate.
• Using web substitutes does not address the protected internal workbook and may introduce irrelevant or inaccurate data.

Question 5

Topic: Draft and Analyze Business Content by Using AI

A sales operations manager used Analyst in Excel to identify Q3 pipeline insights. They now want Microsoft 365 Copilot to turn the insights into one PowerPoint slide for a leadership meeting. The draft prompt is: “Make a slide from these Excel insights.” Which prompt improvement best helps ensure the transferred insights remain accurate, relevant, and formatted for PowerPoint?

Options:

• A. Use only the previous chat and do not check the workbook again.

• B. Make the slide more persuasive and add trends that strengthen the sales story.

• C. Include every worksheet detail so leaders can decide what matters.

• D. Use the Q3 Pipeline workbook, preserve source figures, and create one executive slide with a title, chart, and three bullets.

Best answer: D

Explanation: When moving insights between Microsoft 365 apps, the prompt should name the source, define the target app output, and protect the accuracy of transferred information. In this scenario, PowerPoint needs a concise leadership-ready slide, not a full data dump or a rewritten narrative. Asking Copilot to use the Q3 Pipeline workbook, preserve the source figures, and format the result as a slide with a title, chart, and bullets makes the expected output clearer and easier to verify. The key is to transfer the insight without changing the meaning, overstating the trend, or losing the format needed for the audience.

• Persuasive additions fail because inventing or embellishing trends can make the transferred insight inaccurate.
• Every detail fails because leadership slides need relevant summarized insights, not all worksheet data.
• Previous chat only fails because it removes the source check needed to preserve accurate figures.

Question 6

Topic: Understand Generative AI Fundamentals

A product manager wants to use Microsoft 365 Copilot to help decide whether to continue a pilot program. The accountable business owner must make the final decision.

Draft prompt: “Review the pilot notes and tell us whether to continue the pilot.”

Which prompt improvement best supports informed decision-making without replacing the business owner?

Options:

• A. Ask Copilot for evidence, risks, options, citations, and a decision-ready summary.

• B. Ask Copilot to ignore uncertain data and provide a confident recommendation.

• C. Ask Copilot to summarize only the most positive pilot feedback.

• D. Ask Copilot to make the final go/no-go decision automatically.

Best answer: A

Explanation: For decision-quality work, Microsoft 365 Copilot should help the business owner understand the available information, not replace the owner’s judgment. A stronger prompt asks Copilot to use relevant sources, summarize evidence, identify risks and assumptions, compare options, include citations or source references, and produce an output the owner can review. This supports informed decision-making while preserving human accountability. Prompts that ask Copilot to decide automatically, hide uncertainty, or focus only on favorable evidence increase the risk of over-reliance and poor business judgment.

• Automatic decision-making fails because Copilot should support, not own, accountable business decisions.
• Ignoring uncertainty is unsafe because uncertain or incomplete data should be highlighted for review.
• Positive-only feedback creates bias and omits risks needed for a balanced decision.

Question 7

Topic: Understand Generative AI Fundamentals

A customer success team repeatedly asks Microsoft 365 Copilot for help preparing renewal-risk summaries. They need responses to consistently use the team’s renewal playbook, include the same risk categories, and guide team members with example questions. Which action best satisfies this requirement?

Options:

• A. Rename previous renewal chats for easier searching

• B. Use general chat and ask a new question each time

• C. Create a renewal-risk agent with playbook knowledge and instructions

• D. Draft a one-time summary in Word from the playbook

Best answer: C

Explanation: Agents provide a more focused experience than general chat when a team has a repeated, specific business task. Instead of relying on each user to recreate the right prompt, an agent can be configured with relevant knowledge, task instructions, capabilities, and suggested prompts. In this scenario, the renewal playbook and consistent risk categories should be built into the experience so team members get more consistent guidance for the same type of work. General chat is useful for broad, flexible questions, but it does not package the task context and guidance in the same reusable way.

• General chat is flexible, but it depends on each user restating the playbook context and categories correctly.
• Renaming chats helps find past conversations, but it does not create a focused experience for future renewal summaries.
• One-time drafting may produce a useful document, but it does not guide repeated team use across similar cases.

Question 8

Topic: Manage Prompts and Conversations by Using AI

A sales operations manager is using Microsoft 365 Copilot in PowerPoint to create a 5-slide quarterly pipeline update for regional leaders. The deck must use only the approved Excel pipeline workbook and last quarter’s Word summary, avoid customer-identifying details, and include a short list of assumptions for human review. What is the best business action?

Options:

• A. Create a custom agent before drafting the deck.

• B. Ask Copilot to “make a good pipeline presentation.”

• C. Prompt Copilot with sources, audience, format, privacy limits, and review needs.

• D. Paste customer names from CRM into the prompt for accuracy.

Best answer: C

Explanation: An effective workplace prompt is specific enough for Copilot to produce the intended business outcome. In this scenario, the manager should identify the audience, ask for a 5-slide quarterly pipeline update, reference the approved Excel workbook and Word summary, set privacy limits to exclude customer-identifying details, and request assumptions for human review. This gives Copilot relevant context and guardrails while keeping sensitive information protected. A broad request may produce a generic deck, while adding unnecessary sensitive data or building an agent would not match the immediate drafting need.

• Broad request fails because it does not specify the audience, sources, slide count, privacy limits, or review expectations.
• Sensitive details fail because customer-identifying information is not needed and conflicts with the privacy constraint.
• Custom agent fails because a one-time deck draft can be handled with a well-formed Copilot prompt.

Question 9

Topic: Manage Prompts and Conversations by Using AI

A sales operations manager wants Microsoft 365 Copilot to draft a one-page executive summary of last quarter’s pipeline performance. The summary must be accurate, use approved internal numbers, and avoid exposing individual deal notes. Which resource set should the manager reference in the prompt?

Options:

• A. A prior quarter summary and general web sales benchmarks

• B. Approved Excel pipeline report and finalized quarterly sales deck

• C. All CRM exports, including individual opportunity notes

• D. No files; ask Copilot to infer the trends from memory

Best answer: B

Explanation: The best resource set for Copilot should match the requested output and include the most relevant, authoritative, and appropriately scoped sources. For an executive summary of pipeline performance, approved internal reporting and a finalized sales deck provide reliable figures and business context. They also reduce the chance of including sensitive deal-level details that are not needed for leadership. Adding broad or raw sources can make the response less focused and may increase privacy or accuracy risk. A good prompt should name the task, attach or reference the approved sources, and ask Copilot to cite or reflect the provided data rather than guess.

• Raw CRM exports may contain sensitive individual deal notes that are unnecessary for an executive summary.
• Old and web sources do not meet the need for current, approved internal numbers.
• No referenced files invites unsupported assumptions instead of grounding the summary in verified business data.

Question 10

Topic: Draft and Analyze Business Content by Using AI

A product manager uses Microsoft 365 Copilot to draft weekly status updates for executives. They want Copilot to consistently use a concise tone, include risks before next steps, and avoid technical jargon in future drafts without restating these preferences every time. Which action best satisfies this requirement?

Options:

• A. Create a new agent for executive reporting

• B. Attach the latest project plan to each prompt

• C. Save the current status-update prompt

• D. Add the preferences to Copilot instructions

Best answer: D

Explanation: Copilot instructions are best for explicit, recurring guidance about how responses should be written or structured. In this scenario, the user wants consistent personalization across future drafts: concise tone, risk-first structure, and less technical language. Copilot memory can help personalize responses based on remembered context and preferences, but when the user has clear writing rules, adding them to instructions is the most direct action. Referencing a project plan helps with facts for one task, and saving a prompt helps reuse a request, but neither automatically guides future responses in the same way.

• Project plan reference helps Copilot use current project facts, but it does not store recurring style and structure preferences.
• Saved prompt supports reuse of a specific request, but the user still must choose or run that prompt again.
• New agent may help with repeatable team workflows, but it is unnecessary for a personal drafting preference.

Question 11

Topic: Understand Generative AI Fundamentals

A human resources team wants Microsoft 365 Copilot to help managers get consistent answers about onboarding procedures. The answers must use only approved HR policy documents, avoid exposing individual employee records, and be reusable by several HR coordinators in Microsoft Teams. What is the best business action?

Options:

• A. Build a custom Azure AI application for HR questions

• B. Create an agent from a template with approved HR knowledge and instructions

• C. Ask each coordinator to use their own Copilot chat prompt

• D. Add employee personnel files as agent knowledge sources

Best answer: B

Explanation: For a repeatable business function, an agent is better than separate ad hoc chats because it can be configured with shared knowledge, instructions, capabilities, and suggested prompts. In this scenario, the HR team needs consistent onboarding answers from approved policy documents and reuse across coordinators. Creating an agent from a template keeps the solution at business-user depth while allowing the team to scope the agent to the right HR knowledge and guidance. Sensitive employee records should not be used as knowledge sources unless there is a clear business need and appropriate access. The key takeaway is to use an agent when a team needs a consistent, reusable Copilot experience for a defined business process.

• Separate chats may produce inconsistent answers because each coordinator could use different wording and context.
• Personnel files create unnecessary sensitive-data exposure and are not needed for general onboarding procedure answers.
• Custom development is unnecessarily complex for a business-user need that can be met with an agent template.

Question 12

Topic: Manage Prompts and Conversations by Using AI

A product manager is using Microsoft 365 Copilot to refine a Teams status update. The update says the release moves from May 10 to May 17 because of a QA defect, the customer pilot is not affected, and leadership must decide by Friday whether to approve overtime budget.

Draft prompt: Make this shorter and executive-friendly.

Which prompt refinement best changes the tone, audience, length, or level of detail while preserving the original business meaning?

Options:

• A. Rewrite it as a customer announcement that promises the release will stay on track.

• B. Create a detailed technical root-cause analysis for the engineering team.

• C. Rewrite for a VP audience in a concise professional tone, using no more than three bullets, and preserve the date change, pilot impact, and Friday budget decision.

• D. Make it more positive and remove anything that could concern leadership.

Best answer: C

Explanation: A strong refinement prompt tells Copilot what to change and what to preserve. In this case, the business goal is not to invent a new message; it is to adapt the same status update for senior leaders. The best refinement names the audience, desired tone, length, and required facts: the new release date, pilot impact, and budget decision. That helps Copilot reduce detail while keeping the business meaning intact.

The key takeaway is that tone and length changes should not alter commitments, risks, dates, or decisions in the original content.

• Removing concerns is unsafe because it could hide material risks or decisions leadership needs.
• Changing the audience to engineering adds detail and changes the communication purpose.
• Promising no delay contradicts the original status update and creates a false commitment.

Question 13

Topic: Draft and Analyze Business Content by Using AI

A product manager is using Microsoft 365 Copilot to create a launch brief. The goal is to start from Copilot-generated content, then let marketing and sales refine the same draft asynchronously.

Draft prompt: “Write a launch brief for Project Orion.”

Which improvement best supports the goal?

Options:

• A. Ask Copilot to summarize all past launches

• B. Create separate private drafts for each team

• C. Reference source files and add the draft to Copilot Pages

• D. Ask Copilot to write the final brief for approval

Best answer: C

Explanation: Copilot Pages is designed for developing shared content with others. In this scenario, the team does not only need a one-time Copilot answer; they need a shared place where marketing and sales can refine the same launch brief together. A stronger prompt should include relevant source files or context, ask Copilot for the launch brief, and then move the generated content into Copilot Pages for collaborative editing. This keeps the work anchored in the right references and supports coauthoring instead of scattering feedback across private drafts or email approvals.

• Final-only output misses the need for shared development before approval.
• Separate private drafts make collaboration harder because each team works from a different version.
• All past launches is too broad and does not focus Copilot on the specific launch brief or collaboration goal.

Question 14

Topic: Manage Prompts and Conversations by Using AI

A customer success manager used Microsoft 365 Copilot in Teams to summarize interview notes from three account calls. The summary will be shared with product and sales leaders, must be easy to verify against the original notes, and should not expose customer names. Which follow-up prompt is the best business action?

Options:

• A. “Include the full customer quotes and names so the team has complete context.”

• B. “Rewrite the summary to sound more confident and persuasive for leadership.”

• C. “Create a shareable table with themes, recommended actions, source citations, owners, and anonymized customer references.”

• D. “Create a new agent to summarize all future customer calls automatically.”

Best answer: C

Explanation: A strong follow-up prompt should refine Copilot output for the business need. In this scenario, the team needs a version that is easy to review, cite, reuse, and share, while protecting customer identity. Asking for a structured table with source citations helps reviewers trace conclusions back to the original notes. Adding owners and recommended actions makes the output reusable for follow-up work. Requesting anonymized references addresses the sensitivity constraint. The best prompt improves the existing output instead of simply changing the tone, exposing raw sensitive details, or adding unnecessary complexity.

• Tone-only revision may make the summary read better, but it does not add citations, structure, or sensitivity controls.
• Full quotes and names creates unnecessary sensitive-data exposure when anonymized references would meet the business need.
• New agent creation is unnecessarily complex for refining one summary and does not directly solve the review and sharing requirement.

Question 15

Topic: Manage Prompts and Conversations by Using AI

A product marketing manager is using Microsoft 365 Copilot to develop a quarterly launch plan. The work spans several chats that include source-linked summaries from Word briefs, Teams meeting notes, and follow-up questions. The manager expects to revisit the reasoning over the next month and keep the material organized for a small planning effort without sharing sensitive budget details broadly. What is the best business action?

Options:

• A. Add the relevant conversation to a notebook

• B. Delete the chats after copying the final plan

• C. Share the full chat history with the whole department

• D. Create a new agent for the launch plan

Best answer: A

Explanation: Adding a conversation to a notebook is useful when Copilot work should be preserved, organized, and revisited as part of an ongoing effort. In this scenario, the manager needs the reasoning, source-linked summaries, meeting context, and follow-up questions over several weeks. A notebook supports continuity by keeping relevant conversation material available for later planning and knowledge organization. Because the content includes sensitive budget details, the manager should keep access appropriate rather than broadly sharing the full chat. Creating a new agent would be unnecessary if the need is organizing an active workstream, not building a reusable assistant experience.

• Deleting chats loses useful reasoning and source context needed for the month-long planning effort.
• Broad sharing ignores the sensitivity constraint because budget details should not be exposed to the whole department.
• Creating an agent is more complex than needed when the immediate goal is preserving and organizing Copilot conversation context.

Question 16

Topic: Manage Prompts and Conversations by Using AI

A customer success manager is preparing a Microsoft 365 Copilot agent for the renewal team. The team needs consistent answers based on approved renewal materials and easy starter prompts for common tasks.

Current agent setup

Which action should the manager take before sharing the agent?

Options:

• A. Use a regular Copilot chat instead of sharing the agent.

• B. Add suggested prompts only, because the agent will learn the team process.

• C. Add specific instructions, approved knowledge, and suggested prompts.

• D. Share the agent now and ask team members to write detailed prompts.

Best answer: C

Explanation: A shared agent should be prepared for repeatable team use, not just created and distributed. In this scenario, the agent lacks all three readiness elements: clear instructions for how it should respond, approved knowledge sources to ground answers, and suggested prompts that guide team members toward common renewal tasks. Adding only one of these elements would leave a gap: the agent might be easy to start but not grounded, or grounded but inconsistent in tone and task handling. The best action is to configure all three before sharing so the team gets consistent, relevant, and easier-to-use support.

• Share as-is fails because detailed user prompts do not replace agent instructions, knowledge, and reusable suggested prompts.
• Prompts only fails because suggested prompts guide use but do not provide approved renewal content or response rules.
• Regular chat fails because the requirement is a reusable team agent with configured guidance and knowledge.

Question 17

Topic: Understand Generative AI Fundamentals

A product marketing manager asks Microsoft 365 Copilot in Word to draft a one-page announcement for an upcoming feature release. Copilot returns a generic draft that does not match the audience or approved messaging.

Draft prompt:

Write an announcement about the new feature release.

Which additional context would best improve the prompt?

Options:

• A. Ask Copilot to make the announcement more exciting.

• B. Reference the approved launch brief, target audience, tone, and one-page format.

• C. Paste customer names and private account issues into the prompt.

• D. Include every product document the team has created.

Best answer: B

Explanation: Copilot responses are strongly affected by the context included in the prompt, such as relevant work files, audience, purpose, constraints, and desired output format. In this scenario, the generic result shows that Copilot lacks approved messaging and audience guidance. Referencing the approved launch brief and specifying the target audience, tone, and length helps Copilot ground the draft in the right business information without adding unnecessary or sensitive content.

The key takeaway is to add focused, relevant context rather than vague style requests or broad document dumps.

• Vague improvement fails because “more exciting” does not supply the missing messaging, audience, or format context.
• Overbroad references fail because adding every product document can introduce irrelevant information and reduce focus.
• Sensitive details fail because private customer information should not be added unless it is necessary, permitted, and appropriate for the task.

Question 18

Topic: Draft and Analyze Business Content by Using AI

A product manager has a Microsoft Teams meeting recap that lists launch decisions and owners. They need to update a Word launch brief so executives can review the decisions without reading the meeting thread. Which Microsoft 365 Copilot action best satisfies the requirement?

Options:

• A. Use Copilot in Word and reference the Teams recap

• B. Create a new agent for launch decisions

• C. Use Analyst to create a spreadsheet of action items

• D. Use Copilot in Teams to summarize the meeting again

Best answer: A

Explanation: When insights from one Microsoft 365 source need to appear in another app, start in the app where the final work product belongs and reference the relevant source. Here, the final deliverable is a Word launch brief, and the useful insight is in the Teams meeting recap. Using Copilot in Word with the meeting recap as context helps convert meeting decisions into document-ready content for executives. A Teams summary may help understand the meeting, but it does not directly update the Word brief.

• Teams-only summary repeats the source insight but does not move it into the Word deliverable.
• Spreadsheet output changes the format to action-item analysis, which is not the requested executive brief.
• New agent is unnecessary for a one-time content workflow between Teams and Word.

Question 19

Topic: Draft and Analyze Business Content by Using AI

A product marketing manager uses Microsoft 365 Copilot every Monday to draft a partner update from current status notes. They want future drafts to consistently use an executive-summary format, a calm business tone, and the approved phrase “release timing change” instead of “launch delay” when appropriate, without retyping these preferences each week. What should the manager do before repeating the task?

Options:

• A. Schedule the prompt to run every Monday without changing preferences.

• B. Add the status notes to a Copilot Page for collaboration.

• C. Delete previous conversations before drafting each update.

• D. Update Copilot instructions with the recurring style and wording preferences.

Best answer: D

Explanation: Copilot instructions are used for standing preferences that should shape future responses, such as tone, structure, formatting, and preferred terminology. In this scenario, the source content changes each week, but the desired writing style and wording rules stay the same. Updating instructions helps Copilot apply those preferences repeatedly without requiring the manager to restate them in every prompt. Memory would be more relevant if Copilot had stored an outdated or incorrect preference that needed review or removal. Scheduling a prompt can help with timing, but it does not by itself define the recurring style requirements.

• Deleting conversations removes chat history but does not create reusable writing guidance.
• Using Copilot Pages helps teams collaborate on generated content, not set personal recurring response preferences.
• Scheduling the prompt addresses when the task runs but omits the required style and wording control.

Question 20

Topic: Draft and Analyze Business Content by Using AI

A product manager has a Word document with detailed release notes and needs Microsoft 365 Copilot to help create a presentation for a 10-minute executive briefing. The slides must focus on business impact, risks, and launch decisions rather than technical details. Which prompt approach best satisfies the requirement?

Options:

• A. Paste the entire document into chat and ask for the best slides possible

• B. Ask Copilot to create a technical training deck from web research

• C. Reference the release notes and specify executive audience, purpose, length, focus areas, and slide format

• D. Ask Copilot to make the release notes shorter and easier to read

Best answer: C

Explanation: To turn business source material into presentation-ready content, the prompt should include the referenced source, target audience, business purpose, desired emphasis, and output expectations. In this scenario, Copilot needs to know that the Word release notes are the source, the audience is executives, the presentation is limited to 10 minutes, and the content should prioritize business impact, risks, and launch decisions. These details help Copilot transform a detailed document into a focused briefing rather than a generic summary or technical deck.

The key takeaway is to guide Copilot with context and constraints, not just ask for “slides.”

• Generic summary fails because shortening the notes does not specify a presentation format, audience, or executive decision focus.
• Wrong audience fails because a technical training deck does not match the executive briefing requirement.
• Unstructured request fails because asking for the “best slides” omits the audience, time limit, and required business focus.

Question 21

Topic: Draft and Analyze Business Content by Using AI

A product team used Copilot for meetings to generate a recap after a sprint planning meeting. The project manager wants to turn the meeting output into clear accountability for follow-up work.

Draft prompt: “Summarize the meeting and tell the team what to do next.”

Which prompt revision would best improve the usefulness of the Copilot meeting output?

Options:

• A. Make the recap shorter and more positive for the whole team.

• B. Create an action table with owners, due dates, decisions, and unresolved items from the recap.

• C. Assign every task to the most senior attendee if no owner is listed.

• D. Combine this recap with all project chats and create a final project plan.

Best answer: B

Explanation: Copilot meeting output is most useful for accountability when it is converted into specific, reviewable follow-up information. A strong prompt should ask for structured actions, owners, due dates, decisions, and unresolved questions based on the meeting recap. If the meeting output does not clearly state an owner or date, Copilot should flag that gap rather than invent details. This helps the project manager confirm responsibilities with the team and track follow-through after the meeting. A prompt focused only on tone or brevity may improve readability, but it does not create accountability.

• Tone-only revision may make the recap easier to read, but it does not identify who owns each follow-up item.
• Invented ownership is unsafe because Copilot should not assign work based on seniority when the meeting did not state an owner.
• Overbroad context may introduce irrelevant or sensitive information and goes beyond using the meeting recap for follow-up accountability.

Question 22

Topic: Manage Prompts and Conversations by Using AI

A customer success lead creates an agent from a template to answer questions using an enterprise renewal playbook and account-escalation notes. The knowledge is useful only to enterprise customer success managers, not the full department. Which action best supports safe team readiness?

Options:

• A. Share the agent only with the enterprise CSM group.

• B. Share the agent with the full department.

• C. Post the source files in a Teams channel.

• D. Remove the knowledge sources before sharing it.

Best answer: A

Explanation: When an agent uses business-specific knowledge, team readiness includes deciding who actually needs that agent and sharing it only with that audience. Microsoft 365 Copilot respects organizational permissions, but broad sharing can still create confusion, irrelevant results, or unnecessary access requests. In this scenario, the enterprise renewal and escalation context is relevant only to enterprise customer success managers, so the safest practical approach is targeted sharing with that group. The key is to match the agent’s availability to the business purpose and audience.

• Broad sharing misses the requirement because the knowledge is not relevant to the full department.
• Removing knowledge makes the agent less useful for the intended enterprise renewal task.
• Posting files shares source material instead of safely reusing the configured agent experience.

Question 23

Topic: Draft and Analyze Business Content by Using AI

A product manager used Microsoft 365 Copilot in Word to create a customer announcement from this prompt: “Write a launch announcement for the new service plan.” The draft includes a launch date, pricing, and the claim “guaranteed 40% faster onboarding,” but the source materials only contain a launch brief, a pricing FAQ, and pilot feedback. Which prompt improvement would best prepare the draft for review before sharing?

Options:

• A. Expand the announcement into a full sales playbook.

• B. Revise using the launch brief and pricing FAQ; cite sources and flag unsupported claims.

• C. Make the announcement more confident and remove cautious language.

• D. Use a friendly tone and keep it under one page.

Best answer: B

Explanation: An AI-generated business document is not ready to share just because it is polished. For review readiness, the prompt should direct Copilot to use the right source materials, preserve accuracy, and identify claims that need verification. In this scenario, the draft contains specific facts and a performance claim that are not fully supported by the available materials. Asking for citations and unsupported-claim flags helps reviewers quickly check the launch date, pricing, and onboarding claim before the content is shared with sales leaders. Tone and length can improve readability, but accuracy and source support are the deciding needs here.

• Confident language can make an unsupported claim sound more credible without making it true.
• Full sales playbook broadens the task beyond preparing the announcement draft for review.
• Friendly and concise may improve style, but it does not address source grounding or verification.

Question 24

Topic: Draft and Analyze Business Content by Using AI

A product manager needs to brief the executive team before a budget meeting. They have a 12-page Word report and a Teams meeting transcript with customer feedback. The executives asked for a one-page management summary with key decisions, risks, and evidence they can verify. What is the best business action?

Options:

• A. Share the full transcript and report with the executives

• B. Use Copilot to create a concise executive summary with source references

• C. Use Copilot to generate recommendations without citing sources

• D. Ask Copilot to rewrite the full report in simpler language

Best answer: B

Explanation: A management summary for executives should be concise, decision-focused, and tied to evidence from the source materials. In this scenario, the product manager should use Microsoft 365 Copilot with the Word report and Teams transcript as context, then request a one-page summary that highlights key decisions, risks, and supporting evidence. Because the executives need to verify the content, the prompt should ask for citations or source references and the user should review them before sharing. The goal is not to produce more volume; it is to reduce the material into trusted, actionable insight.

• Simplifying the full report may improve readability, but it does not meet the one-page executive-summary requirement.
• Skipping citations weakens trust because the executives specifically need evidence they can verify.
• Sharing all source material transfers the work to executives instead of summarizing the decisions and risks for them.

Question 25

Topic: Manage Prompts and Conversations by Using AI

A customer success manager has created a shared Microsoft 365 Copilot agent to help the team prepare renewal briefings. Before sharing it broadly, the manager must confirm that team members understand when and how to use the agent for their daily work. Which readiness step best satisfies this requirement?

Options:

• A. Ask team members to use general Copilot chat instead

• B. Run a team review using typical tasks and suggested prompts

• C. Share the agent link and rely on self-service adoption

• D. Add more knowledge sources before any team discussion

Best answer: B

Explanation: For a shared agent, team readiness means more than making the agent available. The team should understand the agent’s purpose, the work scenarios it supports, how to start with suggested prompts, and where human review is still needed. A short review or pilot using familiar renewal-briefing tasks lets the manager confirm that team members can choose the right moments to use the agent and can interact with it effectively. This also exposes confusing instructions or missing suggested prompts before broader use. Simply adding sources or sharing a link may improve access, but it does not confirm user understanding.

• More knowledge only may improve answer relevance, but it does not verify that users know when or how to use the agent.
• General Copilot chat bypasses the shared agent and fails to prepare the team for the intended reusable experience.
• Self-service sharing gives access, but it does not confirm readiness or uncover confusion before rollout.

Questions 26-50

Question 26

Topic: Understand Generative AI Fundamentals

A finance coordinator wants Microsoft 365 Copilot to help prepare a weekly revenue review. The goal is to analyze an Excel workbook and produce a short management summary.

Draft prompt: “Look at this spreadsheet and tell me what happened.”

Which revised prompt is the best improvement?

Options:

• A. Use Analyst to compare Q2 and Q3 renewal revenue in the referenced workbook by region and customer segment, identify the top three drivers of the decline, cite the workbook fields used, and draft a one-page management summary.

• B. Build a custom forecasting model that predicts next quarter’s renewals from the workbook.

• C. Summarize the spreadsheet in detail and make the answer as comprehensive as possible.

• D. Review all company data and explain why revenue dropped, then create whatever report seems useful.

Best answer: A

Explanation: A strong Copilot prompt for business analysis should state the task, identify the source to use, provide the business context, define the expected output, and include a way to check the result. In this scenario, the user needs a nontechnical Copilot approach: use Analyst or Copilot with the referenced workbook to compare revenue, find drivers, and turn the findings into a management summary. Asking for citations or source fields helps reduce over-reliance and supports review. The best prompt stays within business-user scope and avoids asking Copilot to build technical models or use unspecified data.

• All company data is too broad and may imply access beyond the user’s permissions or relevant workbook.
• Custom forecasting model shifts the task toward technical modeling instead of no-code business analysis.
• Comprehensive summary is vague and does not specify comparison, drivers, format, or verification.

Question 27

Topic: Understand Generative AI Fundamentals

A customer success team repeatedly asks Microsoft 365 Copilot to prepare renewal-risk briefings before account meetings. The briefings must use the team’s approved playbook and current account files, follow a standard risk-scoring format, and offer consistent next-step suggestions for all account managers. What is the best business action?

Options:

• A. Create a shared agent with playbook knowledge and instructions

• B. Use a new Copilot chat for each briefing

• C. Build a custom AI app outside Microsoft 365

• D. Ask Copilot to use public web sources only

Best answer: A

Explanation: An agent is the best fit when a repeated business task needs a defined knowledge base, standard instructions, and consistent capabilities or suggested prompts. In this scenario, the team needs briefings grounded in approved internal materials and account files, with the same risk-scoring format and next-step guidance for multiple users. A shared agent can be configured with the relevant knowledge, instructions, and suggested prompts so account managers get more consistent results without rebuilding the prompt each time. A normal chat can still help with one-off work, but it is less suitable for repeatable, teamwide guidance that depends on specific business context.

• One-off chat misses the need for reusable instructions and consistent formatting across account managers.
• Public web only ignores the approved playbook and current account files needed for relevant briefings.
• Custom app build is unnecessarily complex for a business-user productivity need that an agent can address.

Question 28

Topic: Draft and Analyze Business Content by Using AI

A customer success manager needs to create a new Word document that explains a proposed service-renewal plan to senior finance leaders. The document must be concise, use business language, and focus on cost risk, expected value, and next steps. Which Microsoft 365 Copilot action best meets the requirement?

Options:

• A. Ask Copilot in Word to write a renewal document with no added context

• B. Use Copilot in PowerPoint to create slides for the account team

• C. Ask Copilot in Teams to summarize recent renewal meetings

• D. Prompt Copilot in Word with the audience, purpose, key points, tone, and format

Best answer: D

Explanation: For creating a new business document, the best Copilot action is to start in the authoring app, such as Word, and provide a specific prompt. A strong prompt includes the intended audience, business purpose, relevant points, tone, length, and desired structure. In this scenario, finance leaders need a concise document focused on cost risk, expected value, and next steps, so Copilot needs those constraints to shape the draft appropriately. A generic request may produce usable text, but it is less likely to match the audience or business decision need.

• Meeting summary helps gather background but does not directly create the audience-specific document.
• Slide creation uses the wrong format because the requirement is a Word document, not a presentation.
• Generic drafting omits the audience and business focus, making the output less targeted.

Question 29

Topic: Understand Generative AI Fundamentals

A product manager has a finalized Word launch brief and needs a first draft of a presentation for a sales kickoff. The draft should use the brief as the source and include an organized slide structure that the team can edit. Which Microsoft 365 Copilot experience best supports this task?

Options:

• A. Use Copilot in Word to rewrite the brief as a shorter document.

• B. Use Copilot in Outlook to draft an announcement email from the brief.

• C. Use Copilot in PowerPoint to create a presentation from the Word brief.

• D. Use Copilot for meetings to generate follow-up tasks from the brief.

Best answer: C

Explanation: Microsoft 365 Copilot experiences are optimized for the content type and workflow in the current app. When the required output is an editable presentation, PowerPoint is the right experience, especially when the user already has a Word document that can be used as source context. Copilot in PowerPoint can help turn the brief into a slide structure that the team can refine. Outlook would fit an email, Word would fit a document draft or rewrite, and Copilot for meetings would fit meeting recap or follow-up work.

• Email workflow misses the requirement to create a slide deck, even though the brief could also support an announcement.
• Document rewrite stays in Word and does not produce the presentation format the sales team needs.
• Meeting follow-up applies to meeting outputs, not drafting a presentation from a source document.

Question 30

Topic: Manage Prompts and Conversations by Using AI

A customer success manager created a Microsoft 365 Copilot agent from a template to help the team draft consistent renewal-risk summaries. The agent already has the approved playbook as knowledge, clear instructions, and suggested prompts. Other team members need to use the same reusable assistance for the weekly renewal review process. What should the manager do?

Options:

• A. Save the best summary prompt for personal reuse

• B. Share the agent with the team members

• C. Ask each teammate to search Agent Store

• D. Rename the original Copilot conversation

Best answer: B

Explanation: When a business process benefits from repeatable AI assistance, a configured agent should be shared with the people who need to use it. In this scenario, the agent already contains the process-specific knowledge, instructions, and suggested prompts, so sharing it gives the team a consistent experience for the weekly renewal review. A saved prompt helps one user repeat a request, but it does not provide the same shared agent configuration. Searching Agent Store might find general agents, but it would not distribute this team-specific agent.

• Personal prompt reuse misses the team requirement because it keeps the reusable help centered on one user’s prompt.
• Conversation renaming helps organize chat history but does not make an agent available to teammates.
• Agent Store search may find existing agents, but it does not share the already configured renewal-review agent.

Question 31

Topic: Draft and Analyze Business Content by Using AI

A project lead has just finished a Microsoft Teams status meeting. The team needs a clear follow-up that uses the Copilot meeting output to show decisions, action items, owners, and due dates so that commitments can be tracked. What should the project lead do?

Options:

• A. Ask Copilot in PowerPoint to create an executive slide deck

• B. Save the meeting transcript without adding follow-up details

• C. Create a new agent to monitor future project meetings

• D. Use Copilot for meetings to create and share an action-item recap

Best answer: D

Explanation: Copilot meeting output is most useful for accountability when it is converted into a structured follow-up. After a Teams meeting, the project lead should use Copilot for meetings or the meeting recap to summarize key decisions and list action items with owners and due dates, then share that recap with attendees. This supports a common understanding of who is responsible for each next step and when it is expected. A transcript or general summary may preserve information, but it does not by itself create accountable follow-up.

• Slide deck focus may help brief leaders, but it does not directly track owners, decisions, and due dates for the team.
• Transcript only preserves raw meeting content, but it omits the structured follow-up needed for accountability.
• New agent is unnecessary for a single post-meeting accountability workflow and uses the wrong experience.

Question 32

Topic: Understand Generative AI Fundamentals

A sales operations analyst is using Microsoft 365 Copilot in Excel to prepare a renewal-risk summary for an executive meeting. The workbook contains customer contract terms and a column marked “restricted notes” owned by the Customer Success team. A manager asks the analyst to paste the restricted notes into Copilot chat and include them in a shared Copilot Page because Copilot did not return those details from the normal file references. What is the best business action?

Options:

• A. Paste the notes into Copilot and label the page confidential

• B. Ask the data owner or security reviewer before using the restricted notes

• C. Create a new agent to access the restricted notes

• D. Ask Copilot to infer the missing details from available data

Best answer: B

Explanation: Business users should not use Copilot to bypass data boundaries or expand access to sensitive information. In this scenario, Copilot did not return restricted notes through normal file references, and the notes are owned by another team. The safest action is to pause and involve the data owner or a security reviewer to confirm whether the notes may be used, summarized, and shared in a Copilot Page. Copilot can help draft and analyze content, but business users remain responsible for respecting permissions, sensitivity, and review requirements. The key boundary is not a prompt-writing problem; it is an authorization and data-protection decision.

• Confidential label only fails because labeling a page does not authorize use or sharing of restricted source data.
• Inferring missing details fails because it risks fabrication and does not solve the access or sensitivity issue.
• Creating an agent is unnecessarily complex and still would not remove the need for data-owner approval.

Question 33

Topic: Draft and Analyze Business Content by Using AI

A product manager uses Microsoft 365 Copilot each Friday to draft a leadership update. They want every future update to be concise, risk-first, nontechnical, and include a small table of decisions and owners.

Draft prompt: “Summarize this week’s product work and draft the update.”

Which improvement best handles the recurring preference without confusing it with Copilot memory?

Options:

• A. Set the style as a Copilot instruction and prompt with this week’s sources.

• B. Add every future update preference to the Friday prompt only.

• C. Rely on past chats so Copilot infers the preferred format.

• D. Ask Copilot to remember all weekly product work automatically.

Best answer: A

Explanation: Copilot instructions are for recurring guidance about how you want responses shaped, such as tone, format, audience, and preferred structure. Memory is more about remembered facts or preferences from interactions, but it should not be treated as the main way to control a repeated business output format. In this scenario, the stable preference is the leadership-update style, so it should be captured as an instruction. The weekly prompt should still provide the current sources, timeframe, and task so Copilot has the right context for that specific update.

The key distinction is: use instructions for repeatable response behavior, and use each prompt for the current work item.

• Memorizing all work is overbroad and incorrectly assumes Copilot should automatically know or retain all weekly content.
• Inferring from chats is unreliable because past conversations are not a clear, explicit formatting requirement.
• Prompt-only repetition can work once, but it misses the better recurring-use pattern of setting a reusable instruction.

Question 34

Topic: Draft and Analyze Business Content by Using AI

A product manager uses Microsoft 365 Copilot in Word to create an executive management summary from a project status document. The first draft summarizes milestones and risks but does not show what is assumed, what still needs a decision, or what leaders should do next.

Which prompt improvement would best address the gap?

Options:

• A. Ask for assumptions, open questions, and recommended next steps.

• B. Ask Copilot to include every project detail from the source document.

• C. Ask Copilot to make the summary more confident and concise.

• D. Ask Copilot to remove unresolved issues from the summary.

Best answer: A

Explanation: A management summary for executives should do more than restate status. When the source material contains uncertainty or pending decisions, the prompt should ask Copilot to identify assumptions, open questions, and recommended next steps. This helps leaders see what is known, what still needs clarification, and what action is expected. The summary should remain concise, but it should not hide unresolved issues or blur them into the main narrative.

• More confident wording may improve tone, but it does not surface missing assumptions, decisions, or actions.
• Every project detail creates an overly long report instead of an executive-ready summary.
• Removing unresolved issues is unsafe because leaders need visibility into uncertainties and blockers.

Question 35

Topic: Manage Prompts and Conversations by Using AI

A product manager needs a reusable Microsoft 365 Copilot experience for 30 sales reps. The reps should ask questions about a new product launch, get answers grounded in approved launch documents, see customer-ready wording guidance, and start from suggested prompts. The experience must be shared with the team and not require each rep to attach files manually. Which action best satisfies the requirement?

Options:

• A. Use a meeting recap template for launch-document questions

• B. Create a scheduled prompt that emails launch answers daily

• C. Use Microsoft 365 Copilot Chat and attach the files each time

• D. Create an agent from a knowledge Q&A template and share it

Best answer: D

Explanation: For a repeatable team workflow, an agent created from an appropriate template is the best fit. A knowledge or Q&A-style template can be configured with approved launch documents as agent knowledge, instructions for customer-ready wording, relevant capabilities, and suggested prompts. Sharing the agent lets sales reps use the same guided experience without manually attaching the same files in every chat. The key is matching the template to the task and audience: this is not a one-time chat, a recurring email, or a meeting-focused workflow.

• Manual file attachment misses the requirement for a reusable shared experience that does not depend on each rep adding files.
• Scheduled prompt fits recurring updates, not on-demand Q&A grounded in launch materials.
• Meeting recap template targets meeting outputs, not product-launch knowledge retrieval and guided sales wording.

Question 36

Topic: Draft and Analyze Business Content by Using AI

A product manager stepped away for 20 minutes during a Microsoft Teams meeting. When they return, they need to catch up on the missed discussion without interrupting the meeting, especially any decisions or action items. Which Copilot action best satisfies this requirement?

Options:

• A. Ask Copilot in the meeting what was missed

• B. Use Analyst to review project data

• C. Ask Copilot to draft a follow-up email

• D. Create a Copilot Page for future collaboration

Best answer: A

Explanation: Copilot for meetings is designed to help participants understand meeting content, including catching up when they join late or miss part of the discussion. In this scenario, the user needs an in-meeting recap focused on what happened while they were away, not a new document, email, or data analysis. A useful prompt would ask Copilot to summarize the missed portion and highlight decisions, open questions, and assigned action items. This supports the user immediately without disrupting the live conversation. The key distinction is that Copilot for meetings addresses meeting context directly, while other Microsoft 365 Copilot experiences support related but different workflows.

• Collaboration page is useful for shared follow-up content, but it does not directly catch the user up during the meeting.
• Follow-up email may help after the meeting, but it omits the immediate need to understand the missed discussion.
• Analyst review fits business data analysis, not meeting recap or missed-meeting context.

Question 37

Topic: Manage Prompts and Conversations by Using AI

A customer success manager uses Microsoft 365 Copilot every Monday morning to create a pipeline-risk summary from the same sales tracker and recent customer emails. The summary must be ready before a weekly leadership meeting, use only work content the manager is permitted to access, and be reviewed before it is shared. What is the best business action?

Options:

• A. Schedule the prompt to run weekly and review the output

• B. Create a new agent to generate the summary automatically

• C. Share the prompt with all leadership so each person runs it

• D. Save the prompt and remember to run it manually each Monday

Best answer: A

Explanation: Scheduled prompts are intended for recurring business outputs that are needed on a predictable cadence, such as a weekly summary before a standing meeting. In this scenario, the manager already knows the sources, timing, and output pattern, so scheduling the prompt is more reliable than depending on memory. The prompt should reference the relevant work content and still respect organizational permissions; Copilot does not bypass access controls. Because the summary may influence leadership decisions, the manager should review the result for accuracy, citations, and context before sharing it. Saving the prompt is useful for reuse, but scheduling better matches the recurring Monday requirement.

• Manual reuse is less reliable because it depends on the manager remembering to run the prompt before every meeting.
• Broad sharing does not solve the need for one timely, reviewed summary and may create inconsistent outputs.
• New agent is unnecessarily complex when a recurring scheduled prompt meets the need.

Question 38

Topic: Understand Generative AI Fundamentals

A customer success manager could not attend a Microsoft Teams project meeting. The manager needs a reliable recap of decisions and action items by the end of the day. The meeting included confidential customer details, and the follow-up will be sent only to meeting participants. Which action is the best business use of Copilot?

Options:

• A. Create a new agent to handle all future customer meeting follow-ups.

• B. Ask Copilot in PowerPoint to create slides from memory without checking the meeting record.

• C. Use Copilot in Teams to summarize the meeting and review the recap before sharing it with participants.

• D. Paste the meeting transcript into a public AI tool for a faster summary.

Best answer: C

Explanation: Copilot in Teams is best suited for meeting and collaboration tasks such as recaps, decisions, action items, and follow-up drafting when the needed context is in the Teams meeting. Because the meeting included confidential customer information, the recap should stay within the permitted Microsoft 365 collaboration context and be shared only with the appropriate participants. The manager should still review the output against the meeting record before sending it, because Copilot output can omit details or misinterpret discussion. The key takeaway is to use the Copilot experience closest to the work context, then verify before acting on the output.

• Public AI tool fails because it can expose confidential customer details outside the approved work context.
• PowerPoint from memory fails because it ignores the Teams meeting context and skips verification.
• New agent is unnecessarily complex for a one-meeting recap and follow-up task.

Question 39

Topic: Understand Generative AI Fundamentals

A customer success lead wants account managers to prepare weekly renewal-risk reviews using the same approved playbook. Today each person uses a private Microsoft 365 Copilot chat with this draft prompt:

Tell me which renewals need attention this week.

Which improvement would make this work more valuable for the team?

Options:

• A. Create and share a renewal agent grounded in the approved playbook.

• B. Paste sensitive contract terms into a shared prompt.

• C. Keep the prompt in a private chat history for reuse.

• D. Ask Copilot to review all company renewals without references.

Best answer: A

Explanation: A shared agent is more valuable than an individual chat when multiple team members need to repeat the same task using the same business context, instructions, and approved sources. In this scenario, account managers need consistent renewal-risk reviews based on an approved playbook. Creating and sharing an agent lets the team reuse the same guidance and suggested workflow instead of each person inventing a separate prompt in a private chat. The agent can be grounded in appropriate knowledge and shared with the people who need it, while still respecting organizational permissions and data protection.

• Private chat reuse helps one person, but it does not give the team a consistent shared process.
• No references is too broad and can produce unsupported or inconsistent renewal assessments.
• Sensitive sharing increases data-risk exposure and is not needed to improve team reuse.

Question 40

Topic: Understand Generative AI Fundamentals

A customer success team wants Microsoft 365 Copilot to answer account handoff questions in a consistent style, using the team’s approved playbook and escalation rules. The team also wants everyone to reuse the same starting questions. What should the team do?

Options:

• A. Ask each team member to write a detailed Copilot chat prompt manually

• B. Use Copilot in Outlook to summarize each handoff email separately

• C. Delete previous Copilot conversations after each account handoff

• D. Create an agent from a template with team knowledge, instructions, and suggested prompts

Best answer: D

Explanation: For a business function that needs consistent, repeatable responses, a team should create an agent rather than rely on individual one-off chats. In Microsoft 365 Copilot, an agent can be configured with approved knowledge sources, instructions for tone and behavior, capabilities, and suggested prompts. That combination helps the team get answers grounded in the same playbook and follow the same escalation approach. Suggested prompts also make it easier for everyone to start from the same questions. A detailed chat prompt can help one user in one session, but it does not create a shared, reusable experience for the whole customer success function.

• Manual prompting can improve one response, but it depends on each person recreating the same context and wording.
• Outlook summaries help with individual email review, but they do not provide a reusable team agent.
• Deleting conversations manages history, but it does not improve consistency or encode team guidance.

Question 41

Topic: Manage Prompts and Conversations by Using AI

A customer success team is configuring suggested prompts for a Microsoft 365 Copilot agent. The agent uses the team’s renewal playbook, product FAQ, and customer meeting notes as knowledge. The current suggested prompt is: “Help me with a renewal.”

Which replacement would best help account managers start a productive interaction with the agent?

Options:

• A. “Tell me everything about this customer and what I should do next.”

• B. “Write the final renewal offer for every customer due this quarter.”

• C. “Using the renewal playbook and recent meeting notes, create a renewal-call prep brief for [customer] with risks, opportunities, questions to ask, and source citations.”

• D. “Find private concerns from anyone’s notes and use them to pressure the customer.”

Best answer: C

Explanation: Suggested prompts should make it easy for users to begin with a focused, repeatable task. A strong suggested prompt states the business goal, identifies relevant context or knowledge sources, includes placeholders the user can fill in, and asks for a useful output format. In this scenario, account managers need renewal-call preparation, so the prompt should guide the agent to use the renewal playbook and meeting notes, produce a concise prep brief, and include citations or source references for review. That approach is more productive and safer than asking for broad customer information or unsupported recommendations.

• Too broad fails because asking for everything about a customer does not define the task, sources, or output format.
• Overextended action fails because creating final offers for every customer is too broad and skips account-manager review.
• Unsafe use fails because using private concerns to pressure a customer is inappropriate and not a productive business interaction.

Question 42

Topic: Manage Prompts and Conversations by Using AI

A customer success manager is creating a Microsoft 365 Copilot agent to help account managers prepare renewal plans. The agent must use the company’s approved renewal playbook, current discount guidance, and latest quarterly business review template. It should avoid using unrelated team files or generic web information as its main source. What should the manager do?

Options:

• A. Enable web grounding and ask for renewal best practices

• B. Add the entire customer success site as knowledge

• C. Add the specific approved renewal files as agent knowledge

• D. Create suggested prompts without adding knowledge sources

Best answer: C

Explanation: Agent knowledge should match the business task the agent supports. In this scenario, the agent needs to help with renewal plans using approved internal materials, so the best action is to add the specific playbook, discount guidance, and template as knowledge sources. This keeps the agent focused on current, relevant company content and reduces the chance that it will draw from unrelated files or general web information. Suggested prompts can make the agent easier to use, but they do not replace relevant knowledge sources.

• Web-first grounding misses the requirement to rely on approved internal renewal materials.
• Broad site knowledge may include unrelated or outdated files that are not needed for renewal planning.
• Suggested prompts only improve usability but do not provide the approved source content the agent must use.

Question 43

Topic: Manage Prompts and Conversations by Using AI

A sales manager wants Microsoft 365 Copilot to draft a partner-facing summary of a product launch. The manager can access an internal revenue forecast, but that file is marked confidential and the partner team does not have permission to it. The summary should use source-supported information and avoid exposing restricted data. What should the manager do?

Options:

• A. Reference the confidential forecast because Copilot respects permissions.

• B. Reference only approved, audience-accessible files and request citations.

• C. Paste key forecast numbers into the prompt without referencing the file.

• D. Share a prompt that includes the confidential forecast as context.

Best answer: B

Explanation: A good prompt reference strategy uses the minimum relevant sources needed for the task and considers both the requester’s permissions and the intended audience. Microsoft 365 Copilot can use content the user is allowed to access, but that does not mean the resulting output is safe to share with people who lack permission to the source. For a partner-facing summary, the manager should reference only approved launch materials and other sources the partner team is allowed to see, then ask Copilot for citations or source notes to support review. The key takeaway is to avoid turning permitted personal access into an accidental disclosure.

• Permission confusion fails because the manager’s access does not make confidential forecast details appropriate for partner sharing.
• Pasted sensitive data still exposes restricted information, even if the original file is not referenced.
• Shared confidential context spreads a prompt strategy that depends on restricted material and risks inappropriate reuse.

Question 44

Topic: Understand Generative AI Fundamentals

A human resources manager needs a leadership summary of confidential employee survey themes and an internal compensation policy. The source files are stored in Microsoft 365, only the HR team has permission to them, and the summary must cite sources for review. What is the BEST business action?

Options:

• A. Ask Copilot to use only public web results and exclude internal HR files.

• B. Paste the policy and survey text into a public web search tool for summarization.

• C. Create an agent with all HR files and share it broadly with managers.

• D. Use Microsoft 365 Copilot with permitted HR files, request citations, and review the summary.

Best answer: D

Explanation: Microsoft 365 Copilot is designed for workplace use with organizational data protection and user permissions. In this scenario, the HR content is confidential and already stored in Microsoft 365, so the safer business action is to use Copilot against files the manager is allowed to access, ask for citations, and review the result before sharing. Public web search tools are not the right place to paste sensitive internal content because that can expose organizational information outside approved work boundaries. The key distinction is that Copilot works within the Microsoft 365 work context and permissions, while public web search is not a substitute for protected internal analysis.

• Public web search fails because pasting confidential HR content into an external tool can expose sensitive organizational data.
• Web-only Copilot use fails because it ignores the internal files needed for an accurate leadership summary.
• Broad agent sharing fails because it is unnecessary and could expand access to sensitive HR information beyond the intended audience.

Question 45

Topic: Manage Prompts and Conversations by Using AI

A customer success manager uses Microsoft 365 Copilot every Friday to create the same account-health update from a CRM export, recent Teams meeting notes, and a shared action-items document. The manager wants Copilot to run this request on a recurring schedule so the update is ready before the weekly team review. Which prompt-management action best meets this requirement?

Options:

• A. Schedule the prompt to run weekly

• B. Share the prompt with the team

• C. Save the prompt for later use

• D. Rename the conversation by account name

Best answer: A

Explanation: Scheduled prompts are designed for repeated Copilot requests that need to run at a regular time, such as weekly reports, status updates, or recurring team-review materials. In this scenario, the key requirement is not just reusing the wording of the prompt; it is having the update ready before the weekly meeting without manually starting the request each Friday. Saving a prompt helps with reuse, and sharing a prompt helps teammates use the same prompt, but neither satisfies the recurring schedule requirement by itself.

• Saving only helps avoid rewriting the prompt, but the manager would still need to run it manually each week.
• Sharing only supports team reuse, but it does not make the update appear on a weekly schedule.
• Renaming a conversation helps organize chat history, but it does not automate or reuse the recurring report request.

Question 46

Topic: Manage Prompts and Conversations by Using AI

A customer success manager wants Microsoft 365 Copilot to create a useful business summary for leadership from a Word file named Q4 Customer Feedback.docx and an Excel file named Renewal Risks.xlsx. Their draft prompt is: “Summarize customer feedback and tell me what to do.”

Which prompt improvement would most likely produce useful business content?

Options:

• A. “Make the summary more impressive and persuasive for executives.”

• B. “Search everything you can access and include any customer details that help.”

• C. “Use the two named files to create a one-page leadership summary with top themes, renewal risks, recommended actions, and citations.”

• D. “Write a long analysis of all customer issues without asking for sources.”

Best answer: C

Explanation: Effective prompts give Copilot a clear task, relevant context, source material, audience, constraints, and desired output format. In this scenario, the manager needs leadership-ready business content based on two specific files, so the prompt should name those references and define what the summary must include. Asking for citations also helps the user verify the output against source material before using it for decisions. Vague requests such as “make it impressive” do not guide Copilot toward the needed business outcome, and overly broad requests can introduce irrelevant or sensitive information.

• Vague tone request fails because it asks for style without adding source context, business focus, or structure.
• Overbroad access fails because it may pull in unnecessary or sensitive customer information instead of the intended files.
• No source check fails because a long analysis without citations is harder to verify for leadership use.

Question 47

Topic: Manage Prompts and Conversations by Using AI

A customer success manager uses Microsoft 365 Copilot to turn post-call notes into a consistent account-risk brief with sections for risks, next steps, owner, and due date. The task is repeated for different customers after calls, but not on a fixed schedule. The notes may include confidential customer details. What is the best business action?

Options:

• A. Keep copying the prompt from chat history

• B. Save a reusable prompt template with placeholders

• C. Schedule the prompt to run every Monday

• D. Share the prompt with all employees

Best answer: B

Explanation: A saved prompt is best when the same task or output pattern will be reused later, especially when the user will run it manually with different context each time. In this scenario, the manager needs the same brief structure for different customers, but there is no fixed schedule. A reusable prompt template with placeholders helps standardize the output while avoiding hard-coded confidential customer details. The manager can insert or reference the appropriate notes for each account when needed. Scheduling is better for recurring work at a predictable time, and broad sharing is inappropriate when the prompt may encourage use with confidential customer information.

• Scheduling too soon fails because the task happens after calls and has no fixed cadence.
• Broad sharing fails because confidential customer context should not be exposed or encouraged for unnecessary audiences.
• Chat history copying is inefficient and less reliable than saving a reusable prompt template.

Question 48

Topic: Understand Generative AI Fundamentals

A customer success manager uses Microsoft 365 Copilot to draft a renewal email that includes pricing assumptions and promised support outcomes. The email will be sent to a customer and could affect the renewal decision. Which action best meets responsible AI requirements before sending it?

Options:

• A. Send the email with an AI-generated disclaimer

• B. Verify sources and get human review

• C. Ask Copilot to make the email more persuasive

• D. Use only the most recent chat response

Best answer: B

Explanation: When Copilot output could affect a business decision or customer-facing communication, the safe action is to verify it before using it. Copilot can help draft and summarize, but it may produce fabrications, miss context, or reflect incomplete source material. For claims about pricing, commitments, or support outcomes, the user should check citations or referenced files, confirm facts against approved business sources, and have an accountable person review the message before sending. A disclaimer does not replace verification, and making the draft more persuasive can increase risk if the facts are not confirmed.

• Persuasive rewrite improves tone, but it does not confirm whether pricing or support claims are accurate.
• AI disclaimer may add transparency, but it does not make unsupported customer-facing claims safe to send.
• Recent response only increases over-reliance because recency does not prove accuracy or source support.

Question 49

Topic: Manage Prompts and Conversations by Using AI

An HR business partner uses Microsoft 365 Copilot in Word to draft a one-time briefing for a specific employee relations case. The prompt references a confidential investigation file, the exact request will not be reused, and no team members need the prompt. The draft must be checked before it is used. What is the best business action?

Options:

• A. Schedule it before each HR meeting.

• B. Share it with the HR team.

• C. Save it as a standard template.

• D. Use it once privately and review the draft.

Best answer: D

Explanation: Scheduled and shared prompts are useful when a prompt supports repeated work or consistent team reuse. In this scenario, the request is tied to one confidential case, has no recurring schedule, and does not need collaboration. The safest business action is to use the prompt only for the current task, keep it private, and review the Copilot draft before relying on it. Turning the request into a scheduled or shared prompt would add unnecessary reuse and could expose sensitive context. Saving it as a general template is also poor fit because the prompt is case-specific rather than reusable.

• Scheduling reuse fails because the briefing is not a recurring task.
• Team sharing fails because no team reuse is needed and the prompt references confidential case context.
• Standard template fails because a case-specific request is not appropriate as a reusable prompt.

Question 50

Topic: Manage Prompts and Conversations by Using AI

A customer success manager creates a Microsoft 365 Copilot agent from a template to help account teams prepare for quarterly business reviews. The agent should use approved account-plan documents and CRM export summaries, avoid exposing sensitive renewal-risk notes, and help new team members begin useful conversations without guessing what to ask. What is the best business action?

Options:

• A. Add suggested prompts for approved QBR tasks and safe source use

• B. Build a custom app to replace the agent template

• C. Tell users to ask any question in free-form chat

• D. Include renewal-risk notes in every suggested prompt

Best answer: A

Explanation: Suggested prompts are starter questions or tasks that help users understand how to use an agent effectively. In this scenario, the agent has a clear business purpose: preparing for quarterly business reviews using approved sources while avoiding sensitive renewal-risk details. The best action is to configure suggested prompts that model the right work patterns, such as summarizing an account plan, identifying customer priorities, or drafting a QBR agenda from approved documents. This helps new team members start quickly and reduces the chance that they will ask broad, risky, or off-purpose questions. Suggested prompts should align with the agent’s knowledge, instructions, and sensitivity expectations rather than expose restricted information or add unnecessary complexity.

• Free-form only fails because new users may not know what to ask and could use prompts that do not match the agent’s purpose.
• Sensitive notes fails because suggested prompts should not encourage exposing renewal-risk information that should be protected.
• Custom app is unnecessarily complex when a template-based agent with well-configured suggested prompts meets the business need.

Continue in the web app

Use IT Mastery for interactive Microsoft AB-730 practice with mixed sets, timed mocks, topic drills, explanations, and progress tracking.

Try Microsoft AB-730 on Web

Focused topic pages

• Generative AI Fundamentals
• Prompts and Conversations
• Business Content with AI