Free Microsoft AB-731 Exam Questions

Become Microsoft Certified with updated AB-731 exam questions and correct answers

Page: 1 / 66

Total 330 Questions | Updated On: Mar 05, 2026

Add To Cart

Question 1

An administrator wants to build device profiling rules based on network traffic, but the network session view is not populated with any records. Which two settings can be enabled to gather network session information? (Choose two.)

A : Network traffic polling on any modeled infrastructure device

B : Firewall session polling on modeled FortiGate devices

C : Netflow setting on the FortiNAC-F interfaces

D : Layer 3 polling on the infrastructure devices

Answer: B,C

Question 2

Contoso Bank currently uses several AI systems: a model that predicts which customers are likely to churn, an anomaly detection model for fraud, and a new initiative to auto-draft personalised savings-plan emails for clients.In an executive steering committee, the CIO asks you to clarify what is distinct about the new initiative compared to the existing ML models.Which statement best captures the difference between generative AI and other common enterprise AI models in this context?

A : Generative AI is any AI that runs in the cloud rather than on-premises, while other AI models typically run in data centres you own.

B : Generative AI models learn a probability distribution over the data and can produce new content (such as text, images, or code), whereas most other enterprise ML models focus on predicting labels, scores, or anomalies based on existing features.

C : Generative AI always works only on unstructured text, while other AI models are restricted to structured numerical data and cannot be applied to images or audio.

D : Generative AI replaces rule-based engines entirely, while other AI types still rely on explicitly coded if/then rules for every decision.

Answer: B

Question 3

A telco is training a churn model. Midway through, you discover that:10% of labels are wrong due to a bug in how “churn” was recordedSome important features have inconsistent units across systemsMany recent records are missing key fieldsWhat is the most accurate description of how this data quality will affect the model?

A : The model will be largely unaffected because modern algorithms are explicitly designed to ignore label noise and unit inconsistencies during training.

B : Poor labels, inconsistent units, and missing values can distort the patterns the model learns, reducing accuracy and making outputs harder to trust, especially for borderline customers.

C : Only missing values matter; label errors and unit issues are irrelevant because the optimisation process doesn’t look at them directly.

D : Data quality problems affect only training speed, not model behaviour, so your main risk is longer training time rather than worse predictions.

Answer: B

Question 4

What is the main purpose of an AI adoption plan?

A : To document only technical architecture diagrams for AI solutions.

B : To list every possible AI technology available in the market.

C : To allow experimentation with AI without any approvals or governance.

D : To translate AI strategy into actionable steps, timelines, and responsibilities.

Answer: D

Question 5

Fabrikam is moving from a small Copilot pilot to broad rollout. The AI transformation lead wants an AI champions program that actually sustains adoption, not just a one-time launch event.Which design approach best supports a durable AI champions program?

A : Ask for volunteers once and never meet with them again

B : Select only IT staff so the program stays “technical”

C : Rotate champions every week so everyone has a turn

D : Define a clear role, choose cross-functional champions, provide training, and meet regularly as a community

Answer: D

Page: 1 / 66

Total 330 Questions | Updated On: Mar 05, 2026

Add To Cart