Free Exam AB-731: AI Transformation Leader Practice Questions

Sample Practice Questions

Try these 10 sample questions from the Exam AB-731: AI Transformation Leader practice bank. Each includes the answer and explanation. Continue with the interactive practice above to review the full question set.

1. 1Which capability most directly distinguishes generative AI from an AI system that classifies existing records?

   1. A. Assigning records to predefined categories
   2. B. Producing new content from learned patterns
   3. C. Predicting a numeric value from historical data
   4. D. Detecting unusual records in a transaction set

   Answer: B

   Explanation: Microsoft explains that "generative AI goes further by producing new content—such as text, images, audio, or code." Content generation is the distinguishing capability; classification, prediction, and anomaly detection analyze existing data.

   Reference: What is generative AI? — What is generative AI? / opening paragraph

2. 2A team is editing a written proposal and wants embedded Copilot assistance to rewrite passages for clarity and adjust their tone. Which app experience directly fits?

   1. A. Copilot in Excel
   2. B. Copilot in Teams
   3. C. Copilot in Word
   4. D. Copilot in PowerPoint

   Answer: C

   Explanation: The Word row describes "Draft proposals in minutes, rewrite content for clarity, and adjust tone to match your brand voice." Word is the document-writing experience; Excel, Teams, and PowerPoint focus on other work surfaces.

   Reference: Explore Copilot experiences — Copilot in Microsoft 365 apps / Microsoft Word row

3. 3Which Microsoft responsible AI principle calls for treating similarly situated people similarly and testing to reduce bias?

   1. A. Transparency
   2. B. Fairness
   3. C. Privacy and security
   4. D. Accountability

   Answer: B

   Explanation: Microsoft defines "Fairness: Treat similar people and cases similarly by design, and test to reduce bias." The other principles address intelligibility, information protection, and responsibility rather than this equitable-treatment requirement.

   Reference: Identify guiding principles for responsible AI — Six guiding principles / Fairness

4. 4What do language models use to interpret a natural-language request and generate a relevant response?

   1. A. A fixed lookup table of approved answers for every request
   2. B. A customer database containing a response for each sentence
   3. C. Linguistic and semantic relationships among words and phrases
   4. D. A classification label assigned manually to each possible prompt

   Answer: C

   Explanation: The training states that language models "encapsulate the linguistic and semantic relationships between the words and phrases in a vocabulary." These learned relationships support language interpretation and generation rather than exhaustive predefined responses.

   Reference: What is generative AI? — What is generative AI? / language models paragraph

5. 5Which embedded Copilot experience directly supports analyzing trends and creating visualizations from a spreadsheet?

   1. A. Copilot in Excel
   2. B. Copilot in Outlook
   3. C. Copilot in Word
   4. D. Copilot in Teams

   Answer: A

   Explanation: For Excel, Microsoft lists "Analyze trends, create visualizations, and model scenarios instantly." This is spreadsheet analysis, rather than email, document editing, or meeting assistance.

   Reference: Explore Copilot experiences — Copilot in Microsoft 365 apps / Microsoft Excel row

6. 6An AI solution must behave consistently and respond safely when unexpected conditions occur. Which Microsoft responsible AI principle directly addresses this requirement?

   1. A. Inclusiveness
   2. B. Transparency
   3. C. Accountability
   4. D. Reliability and safety

   Answer: D

   Explanation: Microsoft states, "Reliability and safety: Build systems that perform consistently and fail safely, even in unexpected conditions." This principle addresses dependable behavior and safe failure, rather than accessibility, explanations, or assignment of responsibility.

   Reference: Identify guiding principles for responsible AI — Six guiding principles / Reliability and safety

7. 7An assistant must interpret a customer's spoken request together with a photograph of a damaged product. Which model capability should a leader require?

   1. A. Multimodal processing
   2. B. Text-only summarization
   3. C. Code completion
   4. D. Numeric regression

   Answer: A

   Explanation: The official training describes "multi-modal models can handle text, images, and speech". Combining speech with a photograph requires multiple input modalities, rather than text-only processing, code generation, or a numeric prediction.

   Reference: What is generative AI? — What is generative AI? / Note

8. 8A manager wants embedded AI assistance to turn an initiative's ideas into presentation slides. Which app experience is designed for that deliverable?

   1. A. Copilot in Outlook
   2. B. Copilot in Excel
   3. C. Copilot in Teams
   4. D. Copilot in PowerPoint

   Answer: D

   Explanation: The PowerPoint row states, "Transform ideas into polished presentations in minutes." The intended artifact is a slide presentation, rather than a mailbox response, spreadsheet, or meeting recap.

   Reference: Explore Copilot experiences — Copilot in Microsoft 365 apps / Microsoft PowerPoint row

9. 9Which Microsoft responsible AI principle directly calls for protecting data throughout its lifecycle and using least privilege?

   1. A. Privacy and security
   2. B. Inclusiveness
   3. C. Transparency
   4. D. Fairness

   Answer: A

   Explanation: Microsoft states, "Privacy and security: Protect data throughout its lifecycle and design with least privilege and purpose limitation." This principle directly governs information protection and access boundaries.

   Reference: Identify guiding principles for responsible AI — Six guiding principles / Privacy and security

10. 10Which model family is specifically associated with generating images from text prompts?

    1. A. Regression models
    2. B. Classification models
    3. C. Code models
    4. D. Diffusion models

    Answer: D

    Explanation: Under Diffusion Models, Microsoft lists "Generate images from text prompts". Regression predicts numeric values, classification assigns categories, and code models specialize in programming language tasks.

    Reference: Understand generative AI models — Types of generative AI models / Diffusion Models row