Exam AB-731: AI Transformation Leader (AB-731) - Microsoft Actual Exam Questions

A very large generative language model prompted with customer histories to “decide” who is likely to churn, returning a free-form explanation for each user.

A supervised classification model on tabular data (for example, gradient-boosted trees or logistic regression) trained to predict churn probability for each customer.

An unsupervised clustering model that groups customers by similarity without using churn labels, with retention offers sent only to the smallest cluster.

A rule-based system that flags customers randomly until you have enough trial data, and then stops changing once the initial rules are written.

Stand up a small technical team in IT that configures licenses and connectivity, but leaves use-case selection, communication, and training entirely to individual departments.

Form a cross-functional adoption team with business, IT, HR, and communications that owns the use-case backlog, change management, training, and feedback loops in coordination with risk and the AI council.

Ask each department to nominate a “power user” who occasionally joins status calls; keep adoption activity informal so teams don’t feel constrained.

Hire an external consultancy to run adoption workshops for six weeks, then disband the internal group and treat AI usage as purely organic from that point on.

Use Azure AI Search to index and query heterogeneous content with full-text, vector, and hybrid search for your applications and RAG workloads.

Use Azure AI Search to train large language models from scratch in a notebook environment using only your private documents for every scenario in your organisation.

Use Azure AI Search primarily as an identity provider for SaaS apps, replacing Microsoft Entra ID for all sign-in flows.

Use Azure AI Search to host container images and manage Kubernetes clusters through a search-based control plane.

Azure AI Vision analyses images and video, Azure AI Search indexes content, and Azure AI Foundry is mainly a billing portal.

Azure AI Vision adds computer vision, Azure AI Search powers search over indexed data, and Azure AI Foundry is the platform to build and manage AI apps and agents.

Azure AI Vision stores all data, Azure AI Search trains language models, and Azure AI Foundry is a desktop IDE for every workload and region without any configuration changes at all.

Azure AI Vision secures identities, Azure AI Search manages firewalls, and Azure AI Foundry backs up virtual machines.

Azure AI Vision does not secure identities, Azure AI Search does not manage firewalls, and Azure AI Foundry does not back up virtual machines. This choice does not accurately capture the roles of these services in an end-to-end AI solution.

Copilot in Windows is the only version that can create or edit Office documents and manage your work data, while Microsoft 365 Copilot and the Copilot app are limited to answering generic web Questions and never interact with any organisation content.

Copilot in Windows is an OS entry point that can invoke apps, the Copilot app focuses on cross-device chat and agents, and Microsoft 365 Copilot adds rich in- app features grounded in your organisation’s Microsoft Graph work data.

The Copilot app, Copilot in Windows and Microsoft 365 Copilot all share exactly the same capabilities and limits; Microsoft only changes the icon depending on where you open it.

Microsoft 365 Copilot is strictly for consumers on home PCs, while Copilot in Windows and the Copilot app are only for enterprise tenants with Microsoft Entra ID and cannot be used with personal accounts.