Tag: AI

AB-620, AI, Microsoft Certification July 7, 2026

Configure actions and connectors (AB-620 Exam Prep)

This post is a part of the AB-620: Designing and Building Integrated AI Agent Solutions in Copilot Studio Exam Prep Hub.
This topic falls under these sections:
Plan and configure agent solutions (30–35%)
   --> Create and monitor agent flows in Copilot Studio
      --> Configure actions and connectors

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 4 practice tests with 30 questions each available from the hub's main page below the exam topics section.

Introduction

One of the greatest strengths of Microsoft Copilot Studio is its ability to connect AI agents with business applications, enterprise data sources, cloud services, and custom APIs. Rather than simply answering questions, an agent can perform meaningful work on behalf of users by executing actions through connectors.

Actions allow agents to retrieve information, update records, create tickets, send emails, approve requests, trigger workflows, and interact with virtually any connected business system.

Connectors provide the communication bridge between Copilot Studio and external services such as Microsoft 365, Microsoft Dynamics 365, Microsoft Dataverse, Microsoft Fabric, Azure services, Salesforce, ServiceNow, SAP, SQL Server, SharePoint, and thousands of other systems.

For the AB-620 certification exam, you should understand how actions and connectors work together, how to configure them securely, when to use different connector types, and how they support enterprise automation.

What Are Actions?

An action is a task that an AI agent performs in response to user input or workflow logic.

Unlike conversational responses, actions interact with external systems to accomplish business tasks.

Examples include:

Creating a support ticket
Looking up customer information
Sending an email
Updating a CRM record
Creating a calendar event
Starting an approval workflow
Creating a purchase order
Retrieving inventory levels
Executing a REST API
Running a Power Automate flow

Actions transform an AI agent from an information provider into an intelligent business assistant.

What Are Connectors?

A connector is a communication interface that allows Copilot Studio to securely connect to an external service.

Connectors manage:

Authentication
Authorization
API communication
Data formatting
Request execution
Response handling

Instead of writing code to interact with APIs, developers can configure connectors that expose operations as reusable actions.

How Actions and Connectors Work Together

The overall process typically follows these steps:

User submits a request.
Agent understands the intent.
Agent selects an appropriate action.
Action uses a connector.
Connector authenticates with the external service.
External system performs the requested operation.
Results are returned to the agent.
Agent presents the response to the user.

Example:

User:
“Schedule a meeting with Sarah tomorrow.”

Agent:

Identifies scheduling intent.
Uses the Outlook connector.
Creates a calendar event.
Confirms success to the user.

Types of Connectors

Copilot Studio supports several connector types.

Standard Connectors

Standard connectors provide access to many Microsoft and third-party services.

Examples include:

Outlook
SharePoint
OneDrive
Excel
Microsoft Teams
SQL Server
Azure DevOps
Planner

These connectors require little or no custom development.

Premium Connectors

Premium connectors typically connect to enterprise applications.

Examples include:

Salesforce
ServiceNow
SAP
Oracle
Adobe
Azure AI Search
Dynamics 365

Organizations usually require appropriate licensing for premium connectors.

Custom Connectors

When no built-in connector exists, organizations can create custom connectors.

Custom connectors expose:

Internal APIs
Legacy applications
Proprietary services
Industry-specific platforms

They allow Copilot Studio to integrate with virtually any REST-based service.

REST API Tools

Copilot Studio can invoke REST APIs directly.

REST API actions require configuration of:

Endpoint URL
HTTP method
Parameters
Headers
Authentication
Request body
Response schema

REST tools are ideal for integrating with modern web services.

Common Microsoft Connectors

Frequently used connectors include:

Connector	Common Purpose
Outlook	Send emails, manage calendars
Microsoft Teams	Send chat messages and notifications
SharePoint	Read and update documents
Dataverse	Store and retrieve business data
SQL Server	Query relational databases
Azure AI Search	Ground agent responses with enterprise knowledge
Dynamics 365	CRM and ERP operations
Power Automate	Execute workflows
Excel	Read and update spreadsheets
OneDrive	Manage files

Configuring an Action

Creating an action generally involves these steps:

Select the connector.
Authenticate.
Choose the operation.
Configure required parameters.
Configure optional parameters.
Test the action.
Save and publish.

Once configured, the action becomes available within conversations and workflows.

Authentication

Every connector must authenticate before accessing external resources.

Common authentication methods include:

Microsoft Entra ID (Azure AD)
OAuth 2.0
API keys
Basic authentication (legacy scenarios)
Service principals
Managed identities (supported scenarios)

Authentication should follow organizational security policies.

Authorization

Authentication identifies the user or application.

Authorization determines what resources may be accessed.

Examples:

Read customer records
Update inventory
Create invoices
Delete documents

Authorization should always follow the principle of least privilege.

Connection References

Power Platform environments often use connection references.

Connection references:

Separate solution components from connection details.
Simplify deployment between environments.
Reduce manual configuration.
Improve Application Lifecycle Management (ALM).

They are especially useful when moving solutions from development to testing and production.

Using Power Automate Actions

Many Copilot Studio actions invoke Power Automate flows.

Typical scenarios include:

Multi-step approvals
Document generation
Database updates
Email notifications
Scheduled processing
Enterprise orchestration

Power Automate allows complex business logic to remain outside the conversational layer.

Input Parameters

Actions often require user input.

Examples:

Customer ID
Order number
Email address
Product SKU
Invoice number
Employee ID

Copilot Studio can collect missing information during the conversation before invoking the action.

Output Parameters

After execution, actions return information to the agent.

Examples:

Success status
Customer details
Ticket number
Invoice total
Order status
Error message

The agent can use this information to continue the conversation naturally.

Error Handling

Actions should gracefully handle failures.

Examples include:

Authentication failure
Network timeout
Invalid input
API unavailable
Permission denied
Resource not found
Rate limiting

Well-designed agents provide helpful guidance rather than exposing technical errors.

Example:

Instead of:

“HTTP 403 Forbidden.”

Respond:

“I don’t have permission to complete that request. Please contact your administrator or try again using an account with the required permissions.”

Retry Logic

Some failures are temporary.

Examples:

Network interruption
API timeout
Temporary service outage

Retry logic can automatically attempt the operation again before reporting failure.

Security Best Practices

When configuring connectors:

Use secure authentication methods.
Protect credentials.
Avoid hardcoded secrets.
Apply least-privilege permissions.
Use approved enterprise connectors.
Follow Data Loss Prevention (DLP) policies.
Monitor connector usage.
Rotate secrets when required.
Audit access regularly.

Data Loss Prevention (DLP)

Power Platform administrators can define DLP policies that control how connectors are used.

DLP policies help prevent:

Sensitive data leakage
Unauthorized data movement
Mixing business and personal connectors
Compliance violations

For example, an organization may prohibit copying data from Dataverse into personal cloud storage services.

Monitoring Connector Usage

Administrators should monitor:

Failed actions
API latency
Authentication failures
Connector health
Rate limits
Usage frequency
Error rates
Flow performance

Monitoring improves reliability and troubleshooting.

Performance Considerations

Efficient actions improve user experience.

Recommendations include:

Minimize unnecessary API calls.
Request only required data.
Reuse existing connectors.
Cache data when appropriate.
Reduce sequential operations.
Handle pagination efficiently.
Optimize Power Automate flows.

Governance Considerations

Organizations should establish governance for connectors by:

Approving supported connectors
Managing custom connector lifecycle
Reviewing permissions
Monitoring usage
Auditing actions
Documenting integrations
Applying environment strategies
Following change management procedures

Common Design Mistakes

Avoid:

Granting excessive permissions
Ignoring DLP policies
Hardcoding API credentials
Failing to validate user input
Poor error handling
Creating duplicate connectors
Ignoring connection references
Returning technical error messages to users

Best Practices

Choose the simplest connector that satisfies the business requirement.
Prefer built-in connectors when available.
Use custom connectors only when necessary.
Secure authentication using Microsoft Entra ID or OAuth whenever possible.
Validate inputs before invoking actions.
Design user-friendly error messages.
Monitor connector health and usage.
Apply least-privilege access.
Test actions thoroughly before deployment.
Document all external integrations.

Exam Tips

For the AB-620 exam, remember these key points:

Actions perform business operations, while connectors provide access to external systems.
Standard connectors require minimal configuration and support many Microsoft services.
Premium connectors often require additional licensing and connect to enterprise applications.
Custom connectors enable integration with proprietary or unsupported REST APIs.
Power Automate is commonly used to orchestrate complex business processes from Copilot Studio.
Connection references simplify solution deployment across environments.
Always secure connectors using appropriate authentication and least-privilege permissions.
DLP policies govern how connectors can exchange data.
Validate inputs and handle connector errors gracefully.
Monitor connector performance, failures, and usage after deployment.

Practice Exam Questions

Question 1

A Copilot Studio agent needs to create a customer support ticket in an external help desk application. What enables the agent to communicate with that application?

A. A connector

B. A conversation topic

C. A knowledge source

D. A variable

Correct Answer: A

Explanation: Connectors provide the communication interface between Copilot Studio and external services, allowing actions such as creating support tickets.

Question 2

Which statement best describes an action in Copilot Studio?

A. A conversation greeting

B. A reusable business operation performed by an agent

C. A deployment environment

D. A security role

Correct Answer: B

Explanation: Actions execute business tasks such as creating records, sending emails, retrieving data, or calling APIs.

Question 3

An organization must integrate with an internal REST service that has no existing Power Platform connector. Which approach should be used?

A. Use an Adaptive Card

B. Create a Dataverse table

C. Create a custom connector

D. Use conversation variables only

Correct Answer: C

Explanation: Custom connectors expose proprietary or unsupported REST APIs so they can be used as actions within Copilot Studio.

Question 4

Why are connection references recommended when deploying solutions between development and production environments?

A. They eliminate authentication.

B. They permanently embed credentials into solutions.

C. They replace Power Automate.

D. They separate connection details from solution components, simplifying deployment.

Correct Answer: D

Explanation: Connection references improve application lifecycle management by allowing connections to be updated without modifying solution components.

Question 5

Which authentication method is most commonly used with Microsoft cloud services?

A. FTP authentication

B. Microsoft Entra ID (Azure AD)

C. Anonymous authentication

D. Telnet authentication

Correct Answer: B

Explanation: Microsoft Entra ID provides secure identity and access management for Microsoft cloud services and many enterprise integrations.

Question 6

Which Power Platform feature helps prevent sensitive business data from being shared with unauthorized connectors?

A. Conversation history

B. Adaptive Cards

C. Data Loss Prevention (DLP) policies

D. Session variables

Correct Answer: C

Explanation: DLP policies classify connectors and restrict data movement between approved and unapproved services.

Question 7

A connector returns an HTTP timeout while retrieving customer information. What should the agent do?

A. Display the raw HTTP error to the user

B. Retry the operation when appropriate and provide a friendly message if it still fails

C. Ignore the error

D. Delete the connector

Correct Answer: B

Explanation: Temporary failures should be handled gracefully using retry logic and user-friendly error messages.

Question 8

Why should organizations follow the principle of least privilege when configuring connectors?

A. To maximize API usage

B. To eliminate authentication

C. To reduce licensing costs

D. To grant only the permissions required to perform the intended business tasks

Correct Answer: D

Explanation: Least-privilege access minimizes security risks by limiting permissions to only those necessary for the connector’s purpose.

Question 9

Which service is commonly invoked from Copilot Studio to orchestrate multi-step business workflows such as approvals and notifications?

A. Microsoft Paint

B. Microsoft Visio

C. Power Automate

D. Windows Task Scheduler

Correct Answer: C

Explanation: Power Automate is frequently used to execute complex workflows, approvals, notifications, and integrations initiated by Copilot Studio actions.

Question 10

What is the primary benefit of using a built-in standard connector instead of creating a custom connector?

A. It removes all authentication requirements.

B. It generally requires less configuration and maintenance while providing supported integration with common services.

C. It automatically bypasses DLP policies.

D. It can only connect to Microsoft Dataverse.

Correct Answer: B

Explanation: Standard connectors are prebuilt, tested, and supported, reducing development effort and simplifying maintenance compared to custom connectors.

Go to the AB-620 Exam Prep Hub main page

AB-900, Agentic AI, AI Governance, AI Security, Microsoft Certification June 27, 2026

Identify how to configure user access to agents (AB-900 Exam Prep)

This post is a part of the AB-900: Microsoft 365 Copilot and Agent Administration Fundamentals Exam Prep Hub.
This topic falls under these sections:
Perform basic administrative tasks for Copilot and agents (25–30%)
   --> Perform basic administrative tasks for agents
      --> Identify how to configure user access to agents

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 4 practice tests with 30 questions each available from the hub's main page below the exam topics section.

Introduction

In Microsoft 365 Copilot, agents are specialized AI assistants designed to perform focused tasks such as answering domain-specific questions, retrieving organizational knowledge, or executing workflows. Because agents can access organizational data and systems, controlling who can use them and under what conditions is a critical administrative responsibility.

Configuring user access ensures that the right users can interact with the right agents while maintaining security, compliance, and least-privilege principles.

1. What “agent access” means

User access to agents determines:

Which users can discover an agent
Which users can interact with or run an agent
Whether an agent is available organization-wide or restricted to specific groups
Whether external or guest users can use agents (if allowed)

Access is typically controlled through a combination of:

Microsoft 365 identity and access controls
Entra ID (Azure AD) group membership
Copilot and agent-specific policies

2. Key methods to configure access to agents

A. Assigning access via Microsoft Entra ID groups

One of the most common approaches is group-based access control.

Administrators can:

Assign an agent to specific security groups or Microsoft 365 groups
Restrict usage to departments (e.g., HR, Finance, IT)
Manage access at scale without assigning users individually

Benefits:

Scalable management
Easier onboarding/offboarding
Centralized governance

B. Tenant-wide vs scoped availability

Agents can be configured as:

1. Tenant-wide agents

Available to all licensed users in the organization
Used for general productivity scenarios (e.g., company policy assistant)

2. Scoped agents

Limited to specific users or groups
Used for sensitive or department-specific data (e.g., HR policy agent)

C. Role-based access control (RBAC)

Some agent administration actions require specific roles in Microsoft 365 or Entra ID:

Global Administrator
AI Administrator / Copilot Administrator
Service-specific admin roles

RBAC ensures:

Only authorized admins can publish or modify agents
Governance over agent deployment lifecycle

D. Conditional Access policies

Conditional Access can indirectly control agent usage by enforcing:

Device compliance requirements
Multi-factor authentication (MFA)
Location-based restrictions
Risk-based sign-in rules

This ensures that even if a user has access to an agent, they must meet security requirements before using it.

E. Application and permission scopes

Agents may require access to:

Microsoft 365 data (SharePoint, Outlook, Teams)
External connectors or APIs
Graph permissions

Administrators control:

What data the agent can access
Whether consent is required
Whether permissions are user-delegated or app-level

3. Lifecycle considerations for agent access

Provisioning

Define target audience (group or tenant-wide)
Assign initial permissions
Validate compliance requirements

Modification

Update group membership to change access
Adjust policies as organizational needs evolve

Deprovisioning

Remove users or groups when no longer needed
Disable or retire the agent if required
Ensure data access is revoked appropriately

4. Governance best practices

To securely manage agent access:

Use least privilege access (only necessary users/groups)
Prefer group-based assignment over individual assignment
Regularly review agent usage and permissions
Restrict sensitive agents to controlled departments
Monitor access logs for unusual activity
Align with Microsoft Purview policies where applicable

5. Common use cases

HR agent accessible only to HR staff
IT helpdesk agent available to all employees
Finance reporting agent restricted to finance team
Executive summary agent limited to leadership group

6. Key exam takeaway

For AB-900, remember:

Agent access is primarily controlled through Entra ID groups, roles, and policies
Access can be tenant-wide or scoped
Security is enforced through RBAC and Conditional Access
Governance ensures agents are only available to the appropriate users

Practice Exam Questions (10)

1.

What is the most common method used to manage user access to Microsoft 365 agents at scale?

A. Individual user assignment
B. Local device policies
C. Entra ID group-based assignment
D. DNS configuration

Answer: C
Explanation: Entra ID group-based assignment is the scalable and recommended way to manage agent access.

2.

Which configuration limits an agent to only HR department users?

A. Tenant-wide publishing
B. Scoped group assignment
C. Public sharing link
D. Guest user activation

Answer: B
Explanation: Scoped assignment using groups restricts access to specific departments like HR.

3.

Which role is typically required to manage Copilot or agent deployment settings?

A. SharePoint Site Owner
B. Global Administrator
C. Teams Guest User
D. Exchange Recipient User

Answer: B
Explanation: Global Administrators (or similar privileged roles) manage high-level agent deployment settings.

4.

What is the purpose of Conditional Access in relation to agent usage?

A. To increase storage capacity
B. To control data indexing speed
C. To enforce security requirements before access
D. To create new agents automatically

Answer: C
Explanation: Conditional Access ensures users meet security conditions like MFA or device compliance.

5.

What happens when a user is removed from an Entra ID group assigned to an agent?

A. They retain permanent access
B. Their access is automatically revoked
C. The agent is deleted
D. The entire tenant loses access

Answer: B
Explanation: Group membership changes immediately affect access to assigned resources, including agents.

6.

Which access model makes an agent available to all licensed users in a tenant?

A. Scoped access
B. Tenant-wide access
C. External sharing mode
D. Device-based access

Answer: B
Explanation: Tenant-wide access allows all licensed users to use the agent.

7.

Which control helps restrict what data an agent can access?

A. Network firewall rules
B. Permission scopes and Graph permissions
C. Printer access policies
D. Windows registry settings

Answer: B
Explanation: Permission scopes define what data and services an agent can access.

8.

What is a key benefit of using group-based access for agents?

A. It disables auditing
B. It simplifies scalable management
C. It removes the need for authentication
D. It bypasses licensing requirements

Answer: B
Explanation: Group-based access simplifies administration, especially in large organizations.

9.

Which scenario best describes proper agent governance?

A. All users can create unrestricted agents
B. Agents are available without authentication
C. Sensitive agents are limited to specific departments
D. Agents bypass compliance policies

Answer: C
Explanation: Sensitive agents should be restricted to appropriate departments for security and compliance.

10.

What is a recommended best practice when configuring access to agents?

A. Assign access individually to each user
B. Use least privilege access principles
C. Allow anonymous access by default
D. Disable group usage entirely

Answer: B
Explanation: Least privilege ensures users only get the access they need, improving security and governance.

Go to the AB-900 Exam Prep Hub main page

AB-900, AI, AI Governance, AI Security, Microsoft Certification, security June 27, 2026

Discover and manage AI activity by using DSPM for AI (Part 2) (AB-900 Exam Prep)

This post is a part of the AB-900: Microsoft 365 Copilot and Agent Administration Fundamentals Exam Prep Hub.
This topic falls under these sections:
Understand data protection and governance tasks for Microsoft 365 and Copilot (35–40%)
   --> Identify data protection and governance risks for Microsoft 365 and Copilot
      --> Discover and manage AI activity by using DSPM for AI

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 4 practice tests with 30 questions each available from the hub's main page below the exam topics section.

Introduction

In Part 1, you learned how Microsoft Purview Data Security Posture Management (DSPM) for AI helps organizations discover AI activity, identify sensitive data exposure, detect oversharing, and provide visibility into how AI interacts with Microsoft 365 data.

This section (Part 2) focuses on how DSPM for AI helps administrators manage AI-related risks, integrates with other Microsoft security and compliance services, and supports secure AI adoption.

Security Recommendations Generated by DSPM for AI

One of DSPM for AI’s most valuable capabilities is providing actionable security recommendations rather than simply identifying problems.

After analyzing an organization’s AI environment, DSPM highlights areas that should be improved to reduce the likelihood of accidental data exposure or compliance violations.

Examples of recommendations include:

Reduce excessive SharePoint permissions.
Apply sensitivity labels to unclassified confidential files.
Configure Data Loss Prevention (DLP) policies.
Limit external sharing.
Protect highly confidential document libraries.
Enable auditing for AI-related activities.
Improve data governance before expanding AI deployments.

These recommendations help administrators prioritize improvements based on potential business impact and security risk.

Risk Prioritization

Not every security finding represents the same level of risk.

DSPM helps prioritize remediation efforts by evaluating factors such as:

Amount of sensitive data exposed
Number of users with access
Business importance of the data
Existing protection mechanisms
AI usage patterns
Permission inheritance
Regulatory implications

This enables administrators to address the highest-risk issues first.

For example:

Risk	Priority
Public access to executive financial reports	High
Sensitive HR documents lacking labels	High
Marketing presentations shared internally	Medium
Public training documents	Low

Discovering AI-Related Data Exposure

Organizations often ask:

“If we enable Microsoft 365 Copilot today, what sensitive information could users potentially discover?”

DSPM helps answer this question.

It analyzes:

Existing permissions
Data classifications
Sharing configurations
Microsoft Graph relationships
Collaboration patterns

This provides insight into which sensitive data could become more discoverable through AI-assisted searches and summaries.

Remember:

Copilot does not bypass security permissions. It only accesses information that the signed-in user is already authorized to access. DSPM helps identify situations where those permissions may already be too broad.

Remediation Recommendations

After identifying risks, DSPM recommends remediation steps.

Common recommendations include:

Reduce Oversharing

Examples include:

Remove unnecessary SharePoint permissions.
Restrict Microsoft Teams membership.
Remove Everyone access.
Limit guest sharing.

Improve Data Classification

Examples include:

Apply sensitivity labels.
Enable automatic labeling.
Use trainable classifiers.
Configure sensitive information types.

Better classification improves downstream protections across Microsoft Purview.

Strengthen Data Protection Policies

DSPM may recommend:

Creating DLP policies
Encrypting confidential documents
Restricting downloads
Blocking external sharing
Applying retention labels

Review AI Access

Administrators may decide to:

Limit AI rollout to selected departments
Review permissions before enabling Copilot broadly
Reduce access to legacy repositories
Remove stale user accounts

Integration with Microsoft Purview

DSPM for AI does not operate as an isolated product.

Instead, it complements several Microsoft Purview solutions.

Understanding these relationships is important for the AB-900 exam.

Microsoft Purview Information Protection

Information Protection classifies and protects data.

DSPM benefits from these classifications.

For example:

A document labeled:

Highly Confidential
Internal Only
Financial
Legal

helps DSPM understand the sensitivity of AI-accessible content.

Without labels, DSPM has less context when evaluating risk.

Microsoft Purview Data Loss Prevention (DLP)

DLP prevents sensitive information from being shared inappropriately.

DSPM identifies potential risks.

DLP helps enforce policies to prevent those risks from becoming incidents.

Example workflow:

DSPM discovers sensitive payroll files.
DLP prevents external sharing.
Organization reduces AI-related exposure.

Microsoft Purview Insider Risk Management

DSPM identifies risky data exposure.

Insider Risk Management identifies risky user behavior.

Together they help answer two different questions:

DSPM asks:

“What sensitive data could AI access?”

Insider Risk asks:

“Is someone attempting to misuse sensitive data?”

These products complement one another.

Microsoft Purview Activity Explorer

Activity Explorer provides visibility into user interactions with sensitive information.

DSPM can use Activity Explorer insights to better understand:

Sensitive file access
Label usage
DLP events
Data movement

Administrators gain a clearer understanding of how protected information is being used across Microsoft 365.

Microsoft Purview Compliance Manager

Compliance Manager focuses on regulatory compliance.

DSPM focuses on AI data governance.

Together they help organizations:

Reduce compliance risk
Improve governance
Meet regulatory requirements
Protect sensitive information used by AI

Microsoft Defender

Microsoft Defender protects identities, endpoints, applications, and cloud resources.

DSPM complements Defender by focusing specifically on AI-related data risks.

Examples:

Microsoft Defender detects:

Malware
Credential theft
Phishing
Device compromise

DSPM identifies:

Overshared files
AI exposure
Sensitive data visibility
Permission risks

AI Governance Dashboard

DSPM provides dashboards that help administrators understand their organization’s AI posture.

Typical dashboard information includes:

AI adoption trends
Sensitive data exposure
High-risk repositories
Oversharing statistics
AI application inventory
Policy recommendations
Governance posture

Rather than investigating individual files, administrators receive a broad organizational view.

Discovering AI Applications

DSPM helps organizations understand:

Which AI tools are in use
Which departments use them
Adoption trends
AI usage over time

Examples include:

Microsoft 365 Copilot
Microsoft Copilot Chat
Supported third-party AI services

This visibility helps organizations establish AI governance policies.

Investigating AI Risks

Administrators typically investigate findings by asking questions such as:

Which sensitive files are accessible?
Who has access?
Why do they have access?
Is the data properly labeled?
Are permissions appropriate?
Is the data externally shared?
Should additional protection be applied?

DSPM helps surface this information so administrators can make informed decisions.

DLP policies are enabled.

↓

Step 7

Risk is reduced before broader Copilot deployment.

Best Practices

Organizations implementing Microsoft 365 Copilot should follow several best practices.

Review Permissions Before AI Rollout

Avoid enabling Copilot before understanding existing permissions.

Classify Sensitive Data

Use Microsoft Purview Information Protection to classify important documents.

Apply Least Privilege

Users should only have access to information required for their job.

Reduce Oversharing

Review:

SharePoint permissions
Teams memberships
OneDrive sharing
External sharing

Enable DLP

Prevent accidental sharing of confidential information.

Monitor AI Adoption

Understand:

Who uses AI
Which departments use AI
What information AI accesses

Regularly Review Recommendations

DSPM continuously evaluates the environment.

Administrators should regularly review new recommendations as data, permissions, and AI usage evolve.

Licensing Considerations

For the AB-900 exam, you are not expected to memorize licensing details, as licensing can change over time.

However, you should understand these general principles:

DSPM for AI is part of the Microsoft Purview family.
Advanced governance and AI security capabilities may require appropriate Microsoft licensing.
Organizations should verify current licensing requirements before deployment.

Common Exam Scenarios

Incorrect.

DSPM primarily discovers, assesses, and helps reduce AI-related data risks. It is not a prompt-filtering or AI-blocking solution.

Misconception 2

Copilot ignores permissions.

Incorrect.

Copilot always respects the signed-in user’s existing Microsoft 365 permissions.

Misconception 3

DSPM replaces Microsoft Purview DLP.

Incorrect.

DSPM identifies risks, while DLP enforces policies that help prevent inappropriate sharing of sensitive data.

Misconception 4

DSPM replaces Microsoft Defender.

Incorrect.

Defender focuses on threats and attacks, whereas DSPM focuses on AI-related data exposure and governance.

Misconception 5

DSPM automatically fixes security issues.

Incorrect.

DSPM provides visibility, recommendations, and guidance. Administrators remain responsible for implementing changes such as adjusting permissions, applying labels, or configuring policies.

AB-900 Exam Tips

Focus on these key concepts:

Microsoft Purview DSPM for AI is an AI governance and visibility solution.
It helps organizations discover AI usage, identify sensitive data exposure, and reduce AI-related risks.
DSPM does not bypass or modify Microsoft 365 permissions.
It works alongside Information Protection, DLP, Insider Risk Management, Activity Explorer, Compliance Manager, and Microsoft Defender.
One of its primary goals is to identify oversharing before it becomes a business risk.
DSPM provides recommendations, not automatic remediation.
It supports organizations throughout the AI adoption lifecycle by helping them continuously improve their security posture.

Chapter Summary

Microsoft Purview DSPM for AI enables organizations to adopt AI confidently by providing visibility into how AI interacts with organizational data. It discovers AI usage, inventories AI applications, identifies oversharing, evaluates sensitive data exposure, and recommends actions to strengthen governance.

Rather than replacing existing Microsoft Purview or Microsoft Defender capabilities, DSPM for AI enhances them by adding AI-specific insights. It integrates with Information Protection, Data Loss Prevention, Insider Risk Management, Activity Explorer, Compliance Manager, and Microsoft Defender to create a comprehensive approach to AI governance.

For the AB-900 exam, remember that DSPM for AI is fundamentally about discovering, assessing, and managing AI-related data risks. It helps administrators understand where AI could expose sensitive information due to existing permissions and governance gaps, enabling organizations to improve their security posture before and during Microsoft 365 Copilot deployment.

Practice Exam Questions

Question 1

A company plans to deploy Microsoft 365 Copilot across all departments. Before deployment, administrators want to determine whether confidential documents are overly accessible due to existing SharePoint permissions.

Which Microsoft solution should they use?

A. Microsoft Entra Domain Services

B. Microsoft Defender for Endpoint

C. Microsoft Intune

D. Microsoft Purview Data Security Posture Management (DSPM) for AI

Correct Answer: D

Explanation

Microsoft Purview DSPM for AI helps organizations discover overshared content, evaluate AI-related data exposure, and identify permission risks before deploying AI solutions such as Microsoft 365 Copilot.

A is correct because DSPM for AI analyzes permissions and identifies AI-related security risks.
B is incorrect because Defender for Endpoint protects devices.
C is incorrect because Intune manages devices and applications.
D is incorrect because Entra Domain Services provides managed domain services rather than AI governance.

Question 2

An administrator wants to understand which departments are actively using Microsoft 365 Copilot and other approved AI applications.

Which capability best addresses this requirement?

A. Microsoft Purview Information Protection

B. Microsoft Purview DSPM for AI

C. Microsoft Defender for Cloud Apps

D. Microsoft Entra Conditional Access

Correct Answer: B

Explanation

DSPM for AI provides visibility into AI adoption, AI application inventory, and usage trends across the organization.

B is correct because DSPM for AI discovers AI activity and AI adoption.
A classifies and protects data.
C monitors cloud applications but is not specifically designed for AI governance.
D controls authentication conditions.

Question 3

Which statement best describes how Microsoft 365 Copilot accesses organizational data?

A. It bypasses Microsoft 365 permissions when generating responses.

B. It can access all documents stored in Microsoft 365 regardless of permissions.

C. It only accesses content the signed-in user is already authorized to access.

D. It only accesses files created after Copilot was enabled.

Correct Answer: C

Explanation

Copilot respects existing Microsoft 365 permissions. It never bypasses authorization.

C is correct because Copilot only retrieves content the current user can already access.
A and B incorrectly imply that Copilot ignores permissions.
D is incorrect because file creation date is irrelevant.

Question 4

What is the primary purpose of Microsoft Purview DSPM for AI?

A. Prevent all AI-generated responses

B. Replace Microsoft Defender

C. Automatically encrypt all Microsoft 365 data

D. Discover AI activity and identify AI-related data risks

Correct Answer: D

Explanation

DSPM for AI provides visibility into AI usage and helps identify governance and security risks.

D is correct because discovering AI activity and assessing AI-related risks are its primary objectives.
A, B, and C describe capabilities DSPM does not provide.

Question 5

An organization discovers that hundreds of employees can access executive financial reports because of inherited SharePoint permissions.

What type of risk has DSPM for AI identified?

A. Malware infection

B. Oversharing

C. Identity synchronization failure

D. Device compliance failure

Correct Answer: B

Explanation

Oversharing occurs when users have broader access to information than intended.

B is correct because excessive permissions increase AI-related exposure.
A, C, and D are unrelated to data governance.

Question 6

Which Microsoft technology provides much of the contextual relationship information that helps DSPM for AI understand user access to Microsoft 365 content?

A. Microsoft SQL Server

B. Microsoft Defender XDR

C. Microsoft Graph

D. Azure Kubernetes Service

Correct Answer: C

Explanation

Microsoft Graph provides relationships between users, files, emails, Teams, SharePoint, and other Microsoft 365 resources.

C is correct because DSPM uses Microsoft Graph signals to understand data access.
The remaining options do not provide organizational relationship data.

Question 7

Which Microsoft Purview solution works alongside DSPM for AI by preventing inappropriate sharing of sensitive information?

A. Microsoft Purview Data Loss Prevention (DLP)

B. Microsoft Entra ID Protection

C. Microsoft Intune

D. Windows Autopilot

Correct Answer: A

Explanation

DLP enforces policies that prevent sensitive information from being shared improperly.

A is correct because DLP complements DSPM by enforcing protection policies.
B, C, and D serve different purposes.

Question 8

An administrator wants recommendations for reducing AI-related security risks before expanding Microsoft 365 Copilot deployment.

What should they use?

A. Microsoft Defender Antivirus

B. Microsoft Purview DSPM for AI

C. Exchange Online Protection

D. Microsoft Entra Connect

Correct Answer: B

Explanation

DSPM for AI evaluates AI-related risks and recommends improvements such as reducing oversharing, improving data classification, and strengthening governance.

B is correct because providing security recommendations is one of its core capabilities.
The other products address different areas of Microsoft security.

Question 9

Which action would most effectively reduce AI-related data exposure identified by DSPM for AI?

A. Disable Microsoft Teams

B. Increase mailbox quotas

C. Review permissions and apply sensitivity labels to confidential data

D. Upgrade Windows devices

Correct Answer: C

Explanation

Reducing excessive permissions and properly classifying sensitive information significantly reduces AI-related exposure.

C is correct because both permission management and data classification are recommended remediation actions.
A, B, and D do not directly address AI governance.

Question 10

Which statement best summarizes Microsoft’s approach to AI governance with DSPM for AI?

A. DSPM automatically blocks all AI interactions involving confidential information.

B. DSPM replaces Microsoft Purview Information Protection.

C. DSPM eliminates the need for Microsoft Defender.

D. DSPM provides visibility, identifies risks, and recommends actions that help organizations securely adopt AI.

Correct Answer: D

Explanation

Microsoft Purview DSPM for AI is designed to improve organizational AI security posture by discovering AI usage, identifying risks, and recommending governance improvements.

D is correct because it accurately reflects the purpose of DSPM for AI.
A is incorrect because DSPM is primarily a discovery and governance solution rather than an AI-blocking mechanism.
B is incorrect because Information Protection remains responsible for classifying and protecting data.
C is incorrect because Microsoft Defender continues to provide threat protection and complements, rather than is replaced by, DSPM for AI.

Key Takeaways for the AB-900 Exam

After studying this topic, you should be able to:

Explain the purpose of Microsoft Purview DSPM for AI.
Describe how DSPM for AI helps organizations discover and govern AI activity.
Understand that Microsoft 365 Copilot always respects existing user permissions.
Explain the concept of oversharing and why it is a significant AI-related risk.
Describe how Microsoft Graph provides context that enables DSPM for AI to evaluate data access.
Identify how DSPM for AI integrates with Microsoft Purview Information Protection, Data Loss Prevention (DLP), Insider Risk Management, Activity Explorer, Compliance Manager, and Microsoft Defender.
Recognize that DSPM for AI provides visibility, risk assessment, and recommendations, but administrators remain responsible for implementing remediation actions.
Apply DSPM for AI concepts to common AB-900 scenario-based questions involving Microsoft 365 Copilot deployments and AI governance.

These concepts form an important part of the “Identify data protection and governance risks for Microsoft 365 and Copilot” objective and are frequently tested through scenario-based questions that assess your understanding of secure AI adoption and governance.

Go to the AB-900 Exam Prep Hub main page

AB-731, AI, AI Governance, AI Strategy, Microsoft Certification June 12, 2026June 14, 2026

Exam Prep Hub for AB-731: AI Transformation Leader

Welcome to the AB-731: AI Transformation Leader Exam Prep Hub!

Welcome to the one-stop hub with information for preparing for the AB-731: AI Transformation Leader certification exam. The content for this exam helps prepare you to “understand how to recognize opportunities for AI transformation, identify the right AI tools and resources, plan for AI adoption, optimize business processes, guide transformation, and drive innovation by using Microsoft 365 Copilot and Azure AI services”.
Upon successful completion of the exam, you earn the Microsoft Certified: AI Transformation Leader certification.

This hub provides information directly here (topic-by-topic as outlined in the official study guide), links to a number of external resources, tips for preparing for the exam, practice tests, and section questions to help you prepare. Bookmark this page and use it as a guide to ensure that you are fully covering all relevant topics for the AB-731 exam and making use of as many of the resources available as possible.

Audience profile (from Microsoft’s site)



As a candidate for this Microsoft Certification, you should understand how to recognize opportunities for AI transformation, identify the right AI tools and resources, plan for AI adoption, optimize business processes, and drive innovation by using Microsoft 365 Copilot and Azure AI services.

This Certification is designed for business decision-makers at all levels who are responsible for guiding transformation and innovation within their teams or organizations. In this role, you’re expected to demonstrate AI fluency, strategic vision, and the ability to lead AI adoption across teams and functions but are not expected to write any code.

As a candidate for this Certification, you should be able to evaluate AI opportunities, champion responsible AI practices, and align AI investments with business goals. You need experience leading adoption or change management in a business context. You must also be familiar with Microsoft 365 services, Microsoft Foundry, and general AI capabilities.

Skills at a glance (as specified in the official study guide)

Identify the business value of generative AI solutions (35–40%)
Identify benefits, capabilities, and opportunities for Microsoft’s AI apps and services (35–40%)
Identify an implementation and adoption strategy for Microsoft’s AI apps and services (20–25%)

Topic-by-Topic Exam Content

[click a topic link to access the content and practice questions for that topic]

Identify the business value of generative AI solutions (35–40%)

Identify the foundational concepts of generative AI

Identify benefits and capabilities of generative AI solutions

Identify benefits, capabilities, and opportunities for Microsoft’s AI apps and services (35–40%)

Identify benefits and capabilities of Microsoft 365 Copilot and Microsoft Copilot

Identify benefits and capabilities of Foundry Tools

Identify an implementation and adoption strategy for Microsoft’s AI apps and services (20–25%)

Align an AI strategy with Microsoft responsible AI policies

Plan for AI adoption across the organization

AB-731 Practice Exams

Important AB-731 Resources

Link to the free, comprehensive, self-paced course on Microsoft Learn: Drive AI transformation in your organization

https://learn.microsoft.com/en-us/training/courses/ab-731t00

The course has 3 Learning paths:

(1) Explore the business value of generative AI solutions

This learning path has two (2) modules:

(2) Drive business value with AI solutions

This learning path has two (2) modules:

(3) Transform your business with AI

This learning path has four (4) modules:

Good luck to you passing the AB-731 Exam!
However, the more preparation you have, the less luck you will need. 🙂

AB-731, AI, Microsoft Certification June 12, 2026

AB-731 Practice Exam #4 (30 Questions)

This practice exam is a part of the AB-731: AI Transformation Leader Exam Prep Hub.

Question 1 (Scenario-Based)

A global manufacturing company wants to prioritize AI investments. Leadership has identified 25 potential AI use cases across departments.

What should be the FIRST step?

A. Purchase licenses for all users

B. Evaluate each use case based on business value, feasibility, risk, and strategic alignment

C. Deploy AI tools to every department simultaneously

D. Build custom models for every identified opportunity

Answer: B

Explanation

Successful AI transformation begins with prioritization. Organizations should evaluate opportunities according to business impact, implementation complexity, strategic fit, risk, and expected return on investment before committing resources.

Question 2 (Multi-Answer)

A responsible AI review board is evaluating a proposed AI solution.

Which THREE questions should be considered?

A. Could the solution introduce bias?

B. Is user data protected appropriately?

C. Are outcomes explainable to stakeholders?

D. Does the solution maximize token consumption?

E. Does the solution eliminate all human involvement?

Answers: A, B, C

Explanation

Responsible AI reviews focus on fairness, privacy, security, transparency, accountability, and risk mitigation. Token consumption and removing humans from decision-making are not responsible AI objectives.

Question 3 (Single Answer)

An executive wants employees to interact with AI using organizational documents while maintaining existing Microsoft 365 security permissions.

Which Microsoft capability most directly supports this requirement?

A. Microsoft Paint

B. Azure AI Vision

C. Microsoft Graph grounding within Microsoft 365 Copilot

D. Microsoft Defender

Answer: C

Explanation

Microsoft 365 Copilot uses Microsoft Graph to provide context from organizational content while respecting existing permissions and access controls.

Question 4 (Fill in the Blank)

When evaluating AI investments, leaders should focus primarily on measurable __________ rather than technology adoption alone.

A. model parameters

B. infrastructure costs

C. prompt volume

D. business outcomes

Answer: D

Explanation

Business outcomes such as productivity improvements, customer satisfaction, cost reduction, and revenue growth are the primary indicators of AI success.

Question 5 (Match the Answers)

Match the business objective to the most appropriate Microsoft AI capability.

Business Objective	Capability
1. Knowledge retrieval from enterprise content	A. Azure AI Vision
2. Analyze images and extract information	B. Microsoft 365 Copilot
3. Daily productivity assistance	C. Azure AI Search
4. Build custom AI applications	D. Microsoft Foundry

Answers

1 → C
2 → A
3 → B
4 → D

Explanation

Each service addresses a different business need, ranging from search and productivity to custom AI application development.

Question 6 (Scenario-Based)

Question 9 (Scenario-Based)

A healthcare provider wants to use AI to summarize patient records.

Which concern should receive the HIGHEST level of attention?

A. Theme selection

B. Font formatting

C. Privacy and regulatory compliance

D. Presentation templates

Answer: C

Explanation

Healthcare data is highly sensitive and subject to strict privacy and regulatory requirements.

Question 17 (Scenario-Based)

An executive asks whether AI adoption should be measured solely by license assignment rates.

What is the best response?

A. Yes, license assignment is the primary success metric

B. No, business impact and user outcomes should also be measured

C. Yes, adoption metrics replace business KPIs

D. No measurement is required

Answer: B

Explanation

Organizations should measure both adoption and business outcomes to determine whether AI investments deliver value.

Question 18 (Match the Answers)

Match the Responsible AI principle with the corresponding focus.

Principle	Focus
1. Accountability	A. Clear ownership
2. Reliability and Safety	B. Dependable performance
3. Privacy and Security	C. Protecting information
4. Inclusiveness	D. Broad accessibility

Answers

1 → A
2 → B
3 → C
4 → D

Explanation

These pairings align directly with Microsoft’s Responsible AI framework.

Question 19 (Single Answer)

A company wants predictable AI spending across the next fiscal year.

Which purchasing approach may best support this objective?

A. Unlimited experimentation without monitoring

B. Commitment-based or prepaid consumption models

C. Eliminating budgets

D. Monthly spending without forecasting

Answer: B

Explanation

Commitment-based models provide greater spending predictability and budgeting control.

Question 20 (Scenario-Based)

Question 23 (Scenario-Based)

A company wants to identify common barriers that could reduce AI adoption.

Which barrier is often the MOST difficult to overcome?

A. Lack of trust in AI outputs

B. Availability of documentation

C. Number of meeting rooms

D. Internet browser preferences

Answer: A

Explanation

Trust directly affects user willingness to adopt AI tools and integrate them into workflows.

A successful AI adoption team should include business stakeholders, technical experts, and __________ leaders.

A. change management

B. facilities management

C. cafeteria

D. procurement only

Answer: A

Explanation

AI adoption requires organizational change management in addition to technical implementation.

Question 28 (Scenario-Based)

An executive wants evidence that AI investments are producing value.

Which metric provides the STRONGEST evidence?

A. Number of prompts generated

B. Number of licenses purchased

C. Number of AI governance meetings

D. Improvement in targeted business KPIs

Answer: D

Explanation

Business KPIs directly demonstrate whether AI initiatives are achieving intended outcomes.

Question 29 (Single Answer)

Why is inclusiveness important in AI?

A. It ensures AI systems are designed to support diverse users and needs.

B. It eliminates accessibility requirements.

C. It reduces system functionality.

D. It replaces fairness considerations.

Answer: A

Explanation

Inclusiveness promotes accessibility and usability across diverse populations.

Question 30 (Comprehensive Scenario)

A global enterprise has completed several successful AI pilots. Leadership wants to scale AI responsibly while maximizing business value.

Which combination of actions is MOST appropriate?

A. Eliminate governance, automate all decisions, and prioritize rapid deployment

B. Focus only on technology upgrades

C. Establish governance, measure business outcomes, expand adoption programs, and maintain responsible AI oversight

D. Restrict AI usage to executives only

Answer: C

Explanation

Successful enterprise AI transformation requires a balanced approach that combines governance, business value measurement, adoption management, change leadership, and Responsible AI practices. Organizations that scale successfully focus on both innovation and risk management.

Go to the AB-731 Exam Prep Hub main page

AB-731, AI, Microsoft Certification June 12, 2026

AB-731 Practice Exam #3 (30 Questions)

This practice exam is a part of the AB-731: AI Transformation Leader Exam Prep Hub.

Question 1 (Single Answer)

A CEO asks why generative AI initiatives should be tied to business outcomes rather than technology adoption metrics alone.

Which statement best supports this recommendation?

A. AI success should primarily be measured by the number of prompts submitted by employees.

B. AI projects should focus on deploying the newest models available.

C. AI initiatives should be evaluated based on measurable business outcomes such as revenue growth, productivity gains, risk reduction, or customer satisfaction.

D. AI programs should prioritize maximizing model size.

Answer: C

Explanation:
The primary objective of AI transformation is business value creation. Executive leaders should focus on measurable outcomes such as productivity improvements, operational efficiency, customer experience, revenue generation, and risk mitigation. Technology adoption metrics may provide supporting information but are not the primary indicators of success.

Question 2 (Multi-Answer)

A company is evaluating potential generative AI use cases.

Which TWO characteristics indicate a strong candidate for AI investment?

A. The process requires significant manual content creation.

B. The process occurs infrequently and affects only one employee.

C. The process is repetitive and knowledge-intensive.

D. The process cannot tolerate any human review.

E. The process lacks measurable business outcomes.

Answers: A, C

Explanation:
Generative AI delivers significant value when automating or augmenting repetitive knowledge work and content creation tasks. Strong candidates generally affect many users and have measurable business outcomes. Human oversight remains important for most business processes.

Question 3 (Scenario-Based)

A multinational organization wants employees to summarize meetings, draft documents, analyze emails, and retrieve information from Microsoft 365 data.

Which solution best aligns with this requirement?

A. Azure AI Vision

B. Azure AI Search only

C. Azure Machine Learning

D. Microsoft 365 Copilot

Answer: D

Explanation:
Microsoft 365 Copilot integrates with Microsoft 365 applications and organizational content through the Microsoft Graph, helping employees work across Word, Outlook, Teams, Excel, PowerPoint, and other productivity tools.

Question 4 (Single Answer)

An executive team wants AI-generated responses to reference internal company documents while maintaining security permissions.

Which capability primarily enables this?

A. Vector-based retrieval using organizational data

B. Image generation

C. Speech synthesis

D. Computer vision labeling

Answer: A

Explanation:
Retrieval-based architectures using indexed organizational content enable AI systems to ground responses in enterprise knowledge while respecting existing security controls and permissions.

Question 5 (Match the Answers)

Match the Microsoft capability to the most appropriate scenario.

Capability	Scenario
1. Researcher	A. Deep reasoning over structured business data
2. Analyst	B. Multi-step investigation using internal and external information
3. Azure AI Search	C. Enterprise knowledge retrieval
4. Microsoft 365 Copilot	D. Daily productivity assistance

Answers

1 → B
2 → A
3 → C
4 → D

Explanation

Researcher specializes in investigation and synthesis. Analyst focuses on reasoning and analysis. Azure AI Search supports knowledge retrieval, while Microsoft 365 Copilot enhances daily productivity workflows.

Question 6 (Single Answer)

A company wants to create a custom AI experience that uses existing Microsoft 365 Copilot functionality while integrating proprietary business systems.

Which approach should be considered first?

A. Build an entirely new AI platform

B. Replace Microsoft 365

C. Extend Microsoft 365 Copilot

D. Create a manual process

Answer: C

Explanation:
Organizations should typically extend existing capabilities before building entirely new solutions. Copilot extensibility often provides faster time-to-value and lower implementation risk.

Question 7 (Scenario-Based)

A financial institution is evaluating AI opportunities.

Which use case would likely require the highest level of governance oversight?

A. Drafting internal meeting agendas

B. Summarizing project notes

C. Creating social event announcements

D. Assisting with loan approval recommendations

Answer: D

Explanation:
Loan decisions can significantly impact individuals and may involve regulatory, fairness, transparency, and accountability requirements.

Question 8 (Multi-Answer)

Which THREE Microsoft Responsible AI principles are directly concerned with protecting users and ensuring trustworthy outcomes?

A. Reliability and safety

B. Fairness

C. Transparency

D. Revenue optimization

E. Security and privacy

Answers: A, B, E

Explanation:
Reliability and safety, fairness, and privacy/security are foundational principles for trustworthy AI systems. Revenue optimization is not a Responsible AI principle.

Question 9 (Fill in the Blank)

Microsoft recommends that AI-generated content should be reviewed by __________ when the business impact of errors is significant.

A. no one

B. human decision makers

C. model developers only

D. external auditors only

Answer: B

Explanation:
Human oversight remains critical, particularly in high-impact business processes.

Question 10 (Single Answer)

Which business outcome best demonstrates successful AI transformation?

A. Deploying five AI pilots

B. Increasing AI spending

C. Purchasing additional licenses

D. Reducing customer service resolution time by 35%

Answer: D

Explanation:
Business outcomes are the primary measure of AI success. Reduced resolution times represent measurable operational improvement.

Question 11 (Scenario-Based)

An organization wants a generative AI solution capable of processing images, documents, and text within a single workflow.

Which model characteristic is most important?

A. Multimodal capability

B. Smaller context window

C. Traditional relational database support

D. Fixed-output templates

Answer: A

Explanation:
Multimodal models can understand and process multiple content types such as text, images, and documents.

Question 12 (Multi-Answer)

An AI council is being established.

Which TWO responsibilities commonly belong to the council?

A. Defining governance policies

B. Managing every employee prompt

C. Prioritizing AI investments

D. Operating all business applications

Answers: A, C

Explanation:
AI councils provide governance, oversight, prioritization, and strategic direction rather than managing day-to-day operational activities.

Question 13 (Single Answer)

A company wants to identify documents most relevant to a user query across millions of files.

Which Foundry-related capability is most appropriate?

A. Azure AI Search

B. Azure AI Vision

C. Speech Studio

D. Translator

Answer: A

Explanation:
Azure AI Search supports indexing, semantic search, retrieval, and knowledge discovery across large content repositories.

Question 14 (Scenario-Based)

An organization wants to minimize implementation risk while proving AI value.

Which approach is best?

A. Enterprise-wide deployment immediately

B. Replace all existing business processes

C. Launch a targeted pilot with measurable success criteria

D. Delay adoption until competitors finish implementing AI

Answer: C

Explanation:
Pilots allow organizations to validate value, refine governance, and build confidence before scaling.

Question 15 (Single Answer)

Which statement best describes Microsoft Foundry?

A. A replacement for Microsoft 365

B. A platform for building, evaluating, and managing AI solutions and models

C. A cybersecurity monitoring tool

D. A customer relationship management application

Answer: B

Explanation:
Microsoft Foundry provides tools and services for developing, customizing, deploying, and managing AI applications.

Question 16 (Multi-Answer)

Which THREE factors should leaders evaluate when selecting an AI model?

A. Business requirements

B. Cost considerations

C. Performance characteristics

D. Marketing popularity

E. Governance requirements

Answers: A, B, C

Explanation:
Model selection should align with business needs, performance expectations, and budget constraints. Popularity alone is not a reliable criterion.

Question 17 (Single Answer)

What is the primary purpose of transparency in responsible AI?

A. Eliminating governance reviews

B. Explaining how AI systems operate and influence outcomes

C. Increasing token usage

D. Replacing human oversight

Answer: B

Explanation:
Transparency helps users understand AI-generated outputs, limitations, and decision-making processes.

Question 18 (Scenario-Based)

A retailer wants AI to generate personalized marketing content for customers.

What should leadership evaluate first?

A. Whether responsible data usage and privacy requirements are met

B. Whether employees can code in Python

C. Whether every employee owns a GPU

D. Whether all business systems are replaced

Answer: A

Explanation:
Customer data usage introduces privacy, compliance, and governance considerations that must be addressed before deployment.

Question 19 (Single Answer)

Which adoption barrier is most likely to reduce long-term AI success?

A. Strong executive sponsorship

B. Effective governance

C. Lack of user trust

D. Defined business objectives

Answer: C

Explanation:
Without trust, users are less likely to adopt AI tools effectively, reducing realized value.

Question 20 (Match the Answers)

Match each concept to its definition.

Concept	Definition
1. Fairness	A. Protection of information and access
2. Accountability	B. Clear ownership of AI outcomes
3. Security and Privacy	C. Equal treatment across groups
4. Transparency	D. Understanding AI behavior

Answers

1 → C
2 → B
3 → A
4 → D

Explanation

These mappings align directly with Microsoft’s Responsible AI principles.

Question 21 (Single Answer)

Which scenario most strongly supports building a custom AI solution rather than buying an existing one?

A. Standard email summarization

B. General meeting recap generation

C. Highly specialized proprietary workflows requiring unique business logic

D. Drafting common business documents

Answer: C

Explanation:
Custom solutions are most appropriate when unique requirements cannot be adequately met through existing products or extensions.

Question 22 (Multi-Answer)

Which TWO outcomes indicate a successful AI champions program?

A. Increased peer-to-peer knowledge sharing

B. Reduced user engagement

C. Faster adoption of approved AI practices

D. Elimination of governance requirements

Answers: A, C

Explanation:
Champions help accelerate adoption, share best practices, and support organizational learning.

Question 23 (Scenario-Based)

A company wants predictable AI spending for a large, planned deployment.

Which purchasing model may be preferable?

A. Unmanaged consumption only

B. Prepaid or committed capacity approaches

C. Trial subscriptions exclusively

D. Temporary pilot licensing

Answer: B

Explanation:
Prepaid and commitment-based models can improve cost predictability for large-scale deployments.

Question 24 (Single Answer)

What is a key benefit of scalability within Microsoft Foundry?

A. Eliminates governance needs

B. Prevents future upgrades

C. Supports growth from pilots to enterprise deployments

D. Removes the need for monitoring

Answer: C

Explanation:
Scalable platforms help organizations expand AI initiatives without redesigning core architectures.

Question 25 (Fill in the Blank)

The practice of grounding AI responses in approved enterprise knowledge helps improve response __________.

A. randomness

B. reliability

C. ambiguity

D. latency only

Answer: B

Explanation:
Grounding improves factual consistency and reliability by connecting outputs to trusted organizational data.

Question 26 (Single Answer)

An executive sponsor asks why accountability matters in AI governance.

What is the best response?

A. Accountability ensures clear ownership of AI decisions, risks, and outcomes.

B. Accountability removes the need for auditing.

C. Accountability increases model size.

D. Accountability guarantees perfect outputs.

Answer: A

Explanation:
Clear ownership enables organizations to manage risks, governance, and compliance responsibilities.

Question 27 (Scenario-Based)

A healthcare organization plans to use AI-generated recommendations for clinicians.

Which governance action is most important?

A. Eliminating human review

B. Restricting access to executives only

C. Ensuring qualified professionals remain responsible for final decisions

D. Using the largest model available

Answer: C

Explanation:
Healthcare decisions are high-impact and require human oversight and professional accountability.

Question 28 (Multi-Answer)

Which THREE indicators suggest an organization is ready to scale AI adoption?

A. Executive sponsorship

B. Governance framework

C. Demonstrated pilot success

D. Absence of training programs

E. Undefined business goals

Answers: A, B, C

Explanation:
Successful scaling typically requires leadership support, governance structures, and proven pilot outcomes.

Question 29 (Single Answer)

Which statement best distinguishes Microsoft 365 Copilot from Microsoft Foundry?

A. Microsoft 365 Copilot focuses on end-user productivity, while Foundry focuses on building and managing AI solutions.

B. Both products serve identical purposes.

C. Foundry is only for document editing.

D. Microsoft 365 Copilot is only for developers.

Answer: A

Explanation:
Microsoft 365 Copilot is primarily a productivity assistant, whereas Foundry supports AI application development and management.

Question 30 (Scenario-Based)

A global organization wants to evaluate whether an AI solution is creating business value six months after deployment.

Which metric would provide the strongest evidence?

A. Number of prompts entered

B. Number of models deployed

C. Number of governance meetings held

D. Measured improvement in business KPIs tied to the original objectives

Answer: D

Explanation:
Business KPIs provide the clearest evidence that AI investments are delivering intended outcomes. Successful AI transformation is measured by business impact, not technical activity alone.

Go to the AB-731 Exam Prep Hub main page

AB-731, AI, AI Strategy, Generative AI, Microsoft Certification June 12, 2026

Describe the differences between AI models, including fine-tuned and pretrained models (AB-731 Exam Prep)

This post is a part of the AB-731: AI Transformation Leader Exam Prep Hub.
This topic falls under these sections:
Identify the business value of generative AI solutions (35–40%)
   --> Identify the foundational concepts of generative AI
      --> Describe the differences between AI models, including fine-tuned and pretrained models

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 4 practice tests with 30 questions each available from the hub's main page below the exam topics section.

Introduction

Generative AI solutions are powered by AI models that have been trained to recognize patterns, understand language, generate content, and perform a wide variety of tasks. As organizations evaluate AI opportunities, business leaders must understand the different types of AI models available and when each type is appropriate.

One of the most important concepts for the AB-731: AI Transformation Leader exam is understanding the difference between pretrained models and fine-tuned models, as well as how these models fit into broader AI solution strategies.

While technical teams may handle model development and deployment, business leaders must understand the business implications of model selection, including cost, flexibility, performance, governance, and time-to-value.

What Is an AI Model?

An AI model is a system that has learned patterns from data and can use those patterns to perform tasks.

Depending on the model, tasks may include:

Generating text
Answering questions
Creating images
Writing code
Classifying data
Making predictions
Translating languages
Summarizing documents

An AI model can be thought of as the “engine” that powers an AI application.

For example:

Microsoft Copilot uses large AI models to generate responses.
Chatbots use AI models to understand and answer questions.
Image generators use AI models to create pictures from prompts.

Understanding Model Training

AI models learn through a training process.

During training, models analyze large volumes of data and identify patterns, relationships, and structures.

For example, a language model may be trained using:

Books
Articles
Websites
Technical documentation
Publicly available text

After training, the model can generate new content based on what it learned.

The amount of data, computing power, and time required for training can be enormous, especially for modern generative AI systems.

What Is a Pretrained Model?

A pretrained model is an AI model that has already been trained on a large dataset before being made available for use.

Organizations can immediately begin using the model without conducting their own large-scale training.

Characteristics of Pretrained Models

Already trained by the provider
Ready for immediate use
Supports many general-purpose tasks
Requires little or no additional training
Provides rapid deployment

Examples

Many large language models (LLMs) used in enterprise AI solutions are pretrained models.

These models can typically:

Answer questions
Summarize documents
Generate content
Translate languages
Create code

without requiring additional training.

Benefits of Pretrained Models

Faster Time-to-Value

Organizations can begin using the model immediately.

There is no need to spend months collecting and training data.

Example

A company deploys Microsoft Copilot to help employees draft emails and summarize meetings.

The organization benefits from AI capabilities immediately because the underlying model is already trained.

Lower Initial Cost

Training large models from scratch is expensive.

Pretrained models eliminate much of the cost associated with:

Data collection
Model training
Infrastructure
AI expertise

Broad Capabilities

Pretrained models often support many tasks.

Examples include:

Content creation
Summarization
Question answering
Translation
Coding assistance

A single model may address multiple business needs.

Reduced Complexity

Organizations can focus on adoption and business value rather than model development.

Limitations of Pretrained Models

Although pretrained models provide significant advantages, they are not perfect.

Limited Organizational Knowledge

The model may not understand:

Internal policies
Company procedures
Proprietary information
Industry-specific terminology

Generic Responses

Responses may be accurate but lack business-specific context.

Specialized Requirements

Highly regulated or specialized industries may require more tailored behavior.

What Is a Fine-Tuned Model?

A fine-tuned model begins as a pretrained model and then receives additional training using a smaller, targeted dataset.

The goal is to improve performance for a specific task, industry, business process, or domain.

Fine-tuning allows organizations to customize model behavior while leveraging the knowledge already learned during pretraining.

How Fine-Tuning Works

The process generally follows these steps:

Step 1

Start with a pretrained model.

Step 2

Provide additional training data relevant to the desired task.

Step 3

Adjust model parameters based on the specialized data.

Step 4

Deploy the customized model.

Instead of learning everything from scratch, the model builds upon existing knowledge.

Benefits of Fine-Tuned Models

Improved Domain Expertise

Fine-tuned models can better understand:

Industry terminology
Business-specific language
Specialized workflows

Example

A healthcare organization fine-tunes a model using medical documentation and clinical terminology.

The resulting model performs better within healthcare scenarios.

More Consistent Responses

Fine-tuning can help guide the model toward preferred response styles and behaviors.

Example

A company wants all AI-generated customer communications to follow specific branding guidelines.

Fine-tuning can improve consistency.

Better Performance for Specific Tasks

A fine-tuned model often outperforms a general-purpose model when performing specialized tasks.

Examples include:

Legal document analysis
Insurance claims processing
Financial reporting
Industry-specific customer support

Limitations of Fine-Tuned Models

Additional Cost

Fine-tuning requires:

Training resources
Data preparation
Model management

This increases costs compared to simply using a pretrained model.

Data Requirements

Organizations need high-quality training data.

Poor-quality data can reduce model effectiveness.

Ongoing Maintenance

Fine-tuned models may require updates as:

Business processes evolve
Regulations change
New data becomes available

Increased Complexity

Custom models introduce additional governance, testing, and management requirements.

Pretrained vs. Fine-Tuned Models

Characteristic	Pretrained Model	Fine-Tuned Model
Training	Already trained by provider	Additional organization-specific training
Time to deploy	Fast	Longer
Cost	Lower	Higher
Customization	Limited	High
Domain expertise	General	Specialized
Maintenance	Minimal	Greater
Flexibility	Broad tasks	Optimized for specific tasks

Foundation Models

Many generative AI solutions are built on foundation models.

A foundation model is a large AI model trained on enormous amounts of data and capable of supporting many downstream tasks.

Characteristics include:

Large-scale training
Broad capabilities
Adaptability
General-purpose use

Foundation models often serve as the starting point for fine-tuning.

Large Language Models (LLMs)

A Large Language Model (LLM) is a type of foundation model focused on language-related tasks.

Examples of LLM capabilities include:

Writing content
Summarizing information
Translation
Question answering
Conversational interactions

Many Microsoft AI solutions rely on large language models.

Fine-Tuning vs. Retrieval-Augmented Generation (RAG)

Business leaders should understand that fine-tuning is not always required.

Many organizations use Retrieval-Augmented Generation (RAG) instead.

RAG Approach

Rather than retraining the model, RAG:

Retrieves relevant organizational information.
Provides that information to the model.
Generates responses using the retrieved data.

Benefits

Lower cost
Faster implementation
Easier maintenance
Access to current information

Example

An employee asks a question about company policies.

The AI retrieves the latest policy documents and uses them to generate an answer.

The model itself does not need retraining.

For many enterprise scenarios, RAG may be preferable to fine-tuning.

Choosing Between Pretrained and Fine-Tuned Models

Business leaders should evaluate:

Business Requirements

Does the organization need:

General-purpose assistance?
Specialized expertise?

Available Data

Is high-quality domain-specific data available?

Cost Constraints

Can the organization justify customization costs?

Speed of Deployment

How quickly is value needed?

Governance Requirements

What regulatory and compliance considerations apply?

Business Scenarios

Scenario 1: Employee Productivity

Need:

Email drafting
Meeting summaries
Document creation

Best Choice:

Pretrained model

Reason:

General-purpose capabilities are sufficient.

Scenario 2: Industry-Specific Support Assistant

Need:

Specialized terminology
Consistent industry guidance

Best Choice:

Fine-tuned model or RAG-enhanced solution

Reason:

Domain-specific expertise is important.

Scenario 3: Enterprise Knowledge Search

Need:

Access to current internal documents

Best Choice:

RAG solution with a pretrained model

Reason:

Information changes frequently and retraining would be inefficient.

Exam Tips

For the AB-731 exam, remember:

A pretrained model has already been trained and is ready for use.
Fine-tuning adds additional training to customize a pretrained model.
Pretrained models provide faster deployment and lower costs.
Fine-tuned models provide greater specialization and domain expertise.
Foundation models serve as the basis for many generative AI solutions.
Large Language Models (LLMs) are foundation models focused on language tasks.
Fine-tuning is not always necessary; RAG is often a practical alternative.
Business leaders should balance cost, customization, governance, and business value when selecting a model strategy.

Practice Exam Questions

Question 1

A company wants to deploy an AI solution as quickly as possible to help employees draft emails and summarize meetings. Which model approach is most appropriate?

A. Fine-tuned model
B. Pretrained model
C. Custom model trained from scratch
D. Specialized classification model

Answer: B

Explanation: Pretrained models are already trained and can be deployed quickly for general productivity tasks without requiring additional customization.

Question 2

What is the primary purpose of fine-tuning an AI model?

A. Reduce model size
B. Remove training data
C. Improve performance for a specific domain or task
D. Eliminate the need for governance

Answer: C

Explanation: Fine-tuning customizes a pretrained model to perform better within a particular industry, business process, or specialized use case.

Question 3

Which statement best describes a pretrained model?

A. It has already been trained and is ready for use.
B. It requires organization-specific training before deployment.
C. It only supports one task.
D. It contains proprietary company data by default.

Answer: A

Explanation: Pretrained models are trained by the provider and can be used immediately for a variety of general-purpose tasks.

Question 4

A financial services company wants an AI solution that consistently uses industry-specific terminology and follows internal communication standards. Which approach is most likely to help?

A. Disable model training
B. Use only spreadsheets
C. Remove all business data
D. Fine-tune the model

Answer: D

Explanation: Fine-tuning can improve consistency and domain-specific performance by training the model on specialized organizational data.

Question 5

Which characteristic is typically associated with pretrained models?

A. Higher customization
B. Greater maintenance requirements
C. Lower implementation complexity
D. Longer deployment timelines

Answer: C

Explanation: Pretrained models generally require less customization and management, making them easier to implement.

Question 6

What is a foundation model?

A. A database platform for AI applications
B. A large AI model trained on extensive data that supports many tasks
C. A reporting tool used for business intelligence
D. A model that only performs image recognition

Answer: B

Explanation: Foundation models are large-scale models that can support a wide range of downstream AI tasks and applications.

Question 7

Which challenge is most commonly associated with fine-tuned models?

A. Lack of specialization
B. Inability to generate content
C. Additional cost and maintenance requirements
D. Inability to process text

Answer: C

Explanation: Fine-tuning requires additional training, testing, governance, and ongoing management, increasing complexity and cost.

Question 8

An organization needs AI responses based on frequently changing internal policy documents. Which approach may be preferable to fine-tuning?

A. Manual document review only
B. Model retraining every day
C. Predictive analytics
D. Retrieval-Augmented Generation (RAG)

Answer: D

Explanation: RAG retrieves current information at runtime, allowing AI systems to use the latest content without retraining the model.

Question 9

Which factor would most strongly support choosing a pretrained model instead of a fine-tuned model?

A. Need for highly specialized industry knowledge
B. Requirement for maximum customization
C. Desire for rapid deployment and lower cost
D. Availability of extensive proprietary training data

Answer: C

Explanation: Pretrained models are often selected when organizations want quick implementation and lower costs.

Question 10

How does a fine-tuned model typically originate?

A. It is built entirely without training data.
B. It starts as a pretrained model and receives additional targeted training.
C. It is created using only business rules.
D. It is generated automatically by a database.

Answer: B

Explanation: Fine-tuning builds upon an existing pretrained model, allowing it to develop greater expertise in a specific domain or task.

Go to the AB-731 Exam Prep Hub main page

AB-731, AI, AI Strategy, Generative AI, Microsoft Certification June 12, 2026

Describe the Differences Between Generative AI and Other Types of AI (AB-731 Exam Prep)

This post is a part of the AB-731: AI Transformation Leader Exam Prep Hub.
This topic falls under these sections:
Identify the business value of generative AI solutions (35–40%)
   --> Identify the foundational concepts of generative AI
      --> Describe the differences between generative AI and other types of AI

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 4 practice tests with 30 questions each available from the hub's main page below the exam topics section.

Introduction

Artificial Intelligence (AI) has evolved significantly over the past several decades. Organizations now use AI to automate processes, improve decision-making, enhance customer experiences, and create entirely new business opportunities. As AI adoption grows, business leaders must understand the differences between generative AI and other forms of AI because each serves different business purposes and delivers different types of value.

For the AB-731: AI Transformation Leader exam, understanding these distinctions is foundational to evaluating AI opportunities, selecting appropriate solutions, and driving successful AI transformation initiatives.

What Is Artificial Intelligence?

Artificial Intelligence refers to computer systems that can perform tasks that typically require human intelligence. These tasks may include:

Recognizing patterns
Making predictions
Understanding language
Classifying information
Solving problems
Generating content

AI is not a single technology. Instead, it encompasses multiple approaches and capabilities.

Broadly speaking, AI can be divided into two categories:

Traditional (Predictive/Analytical) AI
Generative AI

Traditional AI (Predictive or Analytical AI)

Traditional AI focuses on analyzing existing data to make predictions, classifications, recommendations, or decisions.

Its primary goal is to answer questions such as:

What happened?
What is happening now?
What is likely to happen next?
Which category does this belong to?

Traditional AI learns patterns from historical data and uses those patterns to generate outputs such as predictions or classifications.

Examples

Fraud detection systems
Product recommendation engines
Sales forecasting models
Spam email filtering
Medical image classification
Credit risk assessment

Example Scenario

A bank uses AI to determine whether a credit card transaction is likely fraudulent.

The AI examines:

Transaction amount
Location
Purchase history
Merchant type

The system then classifies the transaction as:

Fraudulent
Not fraudulent

The AI is not creating anything new. It is making a prediction based on existing patterns.

Generative AI

Generative AI goes beyond analyzing data. It creates new content based on patterns learned from large datasets.

Its primary goal is to generate new outputs that resemble human-created content.

Generative AI can produce:

Text
Images
Audio
Video
Code
Summaries
Business documents

Examples

Microsoft Copilot
Large Language Models (LLMs)
AI image generation systems
AI coding assistants
AI-powered content creation tools

Example Scenario

A marketing manager asks Microsoft Copilot to:

Create a marketing campaign for a new product launch.

The AI generates:

Email content
Social media posts
Advertising copy
Campaign ideas

Unlike traditional AI, the system is creating new content rather than classifying or predicting existing data.

Key Difference: Prediction vs. Creation

The simplest distinction is:

Traditional AI	Generative AI
Predicts outcomes	Creates new content
Classifies data	Generates data
Analyzes information	Produces information
Answers “What will happen?”	Answers “What can I create?”
Typically structured outputs	Often natural language outputs

Example

Traditional AI

Input:

Customer purchase history

Output:

Likelihood customer will make another purchase

Generative AI

Input:

Customer profile and product information

Output:

Personalized marketing email

How Traditional AI Works

Traditional AI systems generally follow a supervised learning approach.

The process typically includes:

Collect historical data
Label data
Train a model
Make predictions
Evaluate accuracy

Example

An insurance company may train a model using:

Past claims
Customer demographics
Vehicle information

The model predicts future claim risk.

The output is usually a score, category, or prediction.

How Generative AI Works

Generative AI models are trained on extremely large datasets containing:

Books
Websites
Articles
Images
Code
Documents

The model learns patterns, relationships, structures, and context.

When prompted, it generates new content by predicting the most likely next words, pixels, sounds, or code elements.

Example

Prompt:

Draft a proposal for implementing AI in a customer service department.

Output:

A newly created business proposal tailored to the request.

Foundation Models and Large Language Models

Generative AI is powered by foundation models.

A foundation model is a large AI model trained on enormous amounts of data and capable of supporting many tasks.

Examples include models that can:

Write content
Summarize information
Translate languages
Generate code
Answer questions

A Large Language Model (LLM) is a type of foundation model specialized for language.

Examples include:

GPT models
Models used in Microsoft Copilot
Other enterprise AI language systems

Traditional AI typically uses smaller models trained for specific tasks, while generative AI often relies on large foundation models capable of many tasks.

Deterministic vs. Probabilistic Outputs

Another important distinction is predictability.

Traditional AI

Often produces highly consistent outputs.

Example:

A fraud detection model analyzing the same transaction generally produces the same result.

Generative AI

Produces probabilistic outputs.

Example:

If asked multiple times to create a marketing slogan, the AI may generate different but valid responses.

This flexibility is one reason generative AI is valuable for creativity and content creation.

Data Requirements

Traditional AI

Usually requires:

Structured data
Labeled datasets
Domain-specific training

Examples:

Customer tables
Transaction records
Sensor readings

Generative AI

Uses:

Massive datasets
Structured and unstructured data
Text, images, audio, and code

Examples:

Documents
Books
Emails
Websites
Images

This broader training enables generative AI to perform a wide variety of tasks.

Business Applications of Traditional AI

Organizations commonly use traditional AI for:

Operational Efficiency

Demand forecasting
Inventory management
Route optimization

Risk Management

Fraud detection
Cybersecurity monitoring
Credit scoring

Decision Support

Sales forecasting
Predictive maintenance
Customer churn prediction

The focus is usually on making better business decisions.

Business Applications of Generative AI

Organizations use generative AI to:

Enhance Productivity

Draft emails
Create reports
Generate presentations
Summarize meetings

Improve Customer Experience

Intelligent chatbots
Personalized responses
Conversational assistants

Accelerate Innovation

Product ideation
Content creation
Software development assistance

Knowledge Management

Enterprise search
Document summarization
Knowledge extraction

The focus is often on amplifying human creativity and productivity.

Human Interaction Differences

Traditional AI

Often operates behind the scenes.

Users may not directly interact with the model.

Examples:

Recommendation engines
Risk scoring systems
Automated approval processes

Generative AI

Usually involves direct interaction through prompts and conversations.

Examples:

Microsoft Copilot
AI assistants
Chat-based business applications

Prompt engineering and conversational interaction become important skills.

Benefits of Generative AI Compared to Traditional AI

Generative AI can:

Create content rapidly
Increase employee productivity
Reduce repetitive work
Improve knowledge discovery
Support creativity and innovation
Enable natural language interaction

These capabilities have expanded AI adoption beyond data scientists and technical specialists to everyday business users.

Limitations of Generative AI

Despite its capabilities, generative AI has limitations.

Hallucinations

AI may generate incorrect information that appears credible.

Inconsistent Outputs

Results may vary between prompts.

Governance Requirements

Organizations need policies for:

Data protection
Security
Compliance
Responsible AI

Human Oversight

Generated content often requires review and validation.

Business leaders must understand that generative AI augments human work rather than replacing judgment and accountability.

When to Use Traditional AI vs. Generative AI

Business Need	Best Choice
Fraud detection	Traditional AI
Demand forecasting	Traditional AI
Risk scoring	Traditional AI
Customer segmentation	Traditional AI
Drafting reports	Generative AI
Writing emails	Generative AI
Creating marketing content	Generative AI
Summarizing documents	Generative AI
Conversational assistants	Generative AI
Generating software code	Generative AI

In many organizations, both types of AI work together to deliver business value.

Exam Tips

For the AB-731 exam, remember:

Traditional AI primarily analyzes, predicts, classifies, and recommends.
Generative AI creates new content.
Generative AI is commonly powered by foundation models and large language models.
Traditional AI often works with structured data and task-specific models.
Generative AI works with large-scale structured and unstructured datasets.
Generative AI emphasizes human interaction through prompts and conversations.
Both approaches deliver business value but solve different business problems.

Practice Exam Questions

Question 1

A retail company uses AI to predict which customers are likely to stop purchasing products within the next six months. What type of AI is being used?

A. Generative AI
B. Predictive AI
C. Conversational AI
D. Foundation AI

Answer: B

Explanation: Predicting future customer behavior is a predictive analytics task. The model analyzes historical data and forecasts future outcomes rather than generating new content.

Question 2

Which capability most clearly distinguishes generative AI from traditional AI?

A. Analyzing structured datasets
B. Making classifications
C. Creating new content
D. Detecting patterns

Answer: C

Explanation: The defining characteristic of generative AI is its ability to create new content such as text, images, code, and summaries. Traditional AI primarily analyzes and predicts.

Question 3

A company uses AI to automatically classify incoming support tickets into categories. Which type of AI is primarily being used?

A. Generative AI
B. Foundation AI
C. Traditional AI
D. Conversational AI

Answer: C

Explanation: Ticket categorization is a classification task. Classification is a common traditional AI use case.

Question 4

What is the primary output of a generative AI model?

A. New content based on learned patterns
B. A probability score only
C. A predefined business rule
D. A database query

Answer: A

Explanation: Generative AI creates new outputs such as text, images, code, or summaries based on patterns learned during training.

Question 5

Which business scenario is best suited for generative AI?

A. Fraud detection
B. Inventory forecasting
C. Credit risk scoring
D. Drafting a marketing campaign

Answer: D

Explanation: Creating marketing content requires generating new text and ideas, making it an ideal generative AI use case.

Question 6

How do foundation models differ from many traditional AI models?

A. They only work with structured data.
B. They require no training data.
C. They can support many different tasks after training.
D. They are limited to classification tasks.

Answer: C

Explanation: Foundation models are trained on large datasets and can perform multiple tasks, unlike many traditional AI models that are designed for specific purposes.

Question 7

Which statement about generative AI outputs is most accurate?

A. They are always identical for the same prompt.
B. They are always guaranteed to be correct.
C. They are based solely on business rules.
D. They can vary while still being valid responses.

Answer: D

Explanation: Generative AI is probabilistic and can produce different valid responses to the same prompt.

Question 8

A financial institution uses AI to determine whether a transaction should be flagged as potentially fraudulent. This is an example of:

A. Content generation
B. Predictive classification
C. Creative reasoning
D. Prompt engineering

Answer: B

Explanation: Fraud detection is a classic predictive classification use case where the AI determines the likelihood that a transaction belongs to a fraud category.

Question 9

Which type of data is most commonly associated with traditional AI models?

A. Structured, labeled data
B. Only images
C. Only text documents
D. Randomly generated content

Answer: A

Explanation: Traditional AI frequently relies on structured and labeled datasets for training predictive and classification models.

Question 10

Why do organizations often implement both traditional AI and generative AI?

A. Traditional AI can only be used in research environments.
B. Generative AI eliminates all predictive modeling needs.
C. The two approaches solve different business problems and complement one another.
D. Foundation models require traditional AI to function.

Answer: C

Explanation: Traditional AI excels at prediction and classification, while generative AI excels at content creation and conversational experiences. Together they provide broader business value.

Go to the AB-731 Exam Prep Hub main page

AB-730, AI, Artificial Intelligence (AI), Generative AI, Microsoft Certification June 6, 2026

Understand how to find previous conversations (AB-730 Exam Prep)

This post is a part of the AB-730: AI Business Professional Exam Prep Hub.
This topic falls under these sections:
Manage prompts and conversations by using AI (35–40%)
   --> Manage conversations in Copilot
      --> Understand how to find previous conversations

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

One of the most valuable features of Microsoft 365 Copilot is its ability to maintain conversation history. As users interact with Copilot throughout their workday, they often create summaries, draft documents, analyze data, brainstorm ideas, and ask questions. Rather than starting over each time, users can revisit previous conversations to continue work, retrieve information, review outputs, or refine earlier results.

Understanding how to locate and use previous conversations is an important skill for the AB-730: AI Business Professional exam because it helps improve productivity, supports collaboration, and enables users to build upon prior interactions with AI.

What Are Previous Conversations?

A conversation is an interaction between a user and Copilot that contains:

Prompts submitted by the user
Responses generated by Copilot
Follow-up questions
Revisions and refinements
Referenced files or resources

Over time, users may accumulate many conversations covering different projects, topics, and business activities.

Previous conversations provide a record of these interactions that can be reviewed and reused.

Why Finding Previous Conversations Is Important

Without conversation history, users would need to recreate prompts and repeat work.

Access to previous conversations allows users to:

Resume ongoing work
Reuse successful prompts
Review previous outputs
Verify information
Maintain project continuity
Save time and effort

This makes Copilot a more effective productivity tool.

Common Reasons for Revisiting Conversations

Continuing an Existing Task

A user may begin drafting a report one day and finish it later.

Instead of creating a new conversation, the user can reopen the previous conversation and continue working.

Example:

A marketing manager begins creating a campaign plan on Monday and revisits the conversation on Wednesday to refine the messaging.

Reusing Effective Prompts

Users often discover prompts that consistently produce useful results.

By locating a previous conversation, they can:

Reuse the prompt
Modify the prompt
Share the prompt with others

This reduces the need to recreate successful prompts.

Reviewing Generated Content

Previous conversations can contain valuable outputs such as:

Meeting summaries
Project reports
Business analyses
Draft emails
Presentations
Action plans

Users can revisit these outputs as needed.

Verifying Earlier Work

Users may need to confirm:

What was asked
What Copilot generated
Which files were referenced
What conclusions were reached

Conversation history supports auditing and verification.

Conversation History in Copilot

Microsoft 365 Copilot provides access to prior conversations through conversation history features.

Depending on the Copilot experience and application, users can typically:

View recent conversations
Browse conversation history
Reopen prior chats
Continue existing discussions

The exact interface may vary as Microsoft updates the product, but the underlying concept remains the same.

Benefits of Conversation History

Improved Productivity

Instead of recreating work, users can continue where they left off.

This saves time and effort.

Better Context Retention

Previous conversations contain context that may be useful for future interactions.

For example:

A project discussion may include:

Objectives
Risks
Stakeholders
Action items

Reopening the conversation allows the user to continue working within that context.

Reduced Repetition

Users do not need to repeatedly explain the same background information.

The previous conversation already contains much of the context.

Knowledge Preservation

Conversation history serves as a record of AI-assisted work.

This can be valuable for future reference.

Searching for Previous Conversations

Organizations may accumulate large numbers of conversations over time.

Finding a specific conversation may involve:

Reviewing conversation titles
Browsing recent activity
Searching for keywords
Looking for specific topics or projects

Effective organization helps users locate conversations more quickly.

Naming and Organizing Conversations

Although interfaces vary, users benefit from keeping conversations focused and clearly identifiable.

Examples include:

Q3 Sales Analysis
Marketing Campaign Draft
Executive Meeting Summary
Product Launch Plan

Meaningful names and topics make conversations easier to find later.

Continuing a Previous Conversation

One advantage of locating a previous conversation is the ability to continue it.

Example:

Original prompt:

Summarize the project status and identify key risks.

Several days later, the user reopens the conversation and asks:

Update the analysis using this week’s project data.

The conversation continues instead of starting from scratch.

Previous Conversations and Context

A key exam concept is understanding that previous conversations can provide context.

When continuing an existing conversation:

Prior prompts may influence the discussion.
Earlier outputs may be referenced.
Existing context may improve continuity.

However, users should still verify that the context remains relevant and accurate.

Security and Access Controls

Conversation history remains subject to organizational security policies.

Important exam concepts include:

Security controls continue to apply.
Access permissions remain enforced.
Conversation history does not grant new permissions.
Users can only access information they are authorized to access.

Finding a conversation does not override organizational governance policies.

Data Protection Considerations

Previous conversations may contain references to:

Documents
Emails
Reports
Business data

Organizations should follow established policies regarding:

Data retention
Information governance
Confidentiality
Compliance requirements

Users should avoid sharing sensitive conversation content inappropriately.

Responsible AI Considerations

Even when reviewing previous conversations, users should remember:

AI-generated content may contain errors.
Earlier outputs may become outdated.
Business conditions may have changed.
Human review remains necessary.

Past outputs should not automatically be assumed to be correct.

Conversation History vs. Saved Prompts

These concepts are related but different.

Conversation History

Contains the entire interaction:

Prompts
Responses
Follow-up discussions

Saved Prompt

Contains only the reusable prompt itself.

A saved prompt can be used in many conversations, while conversation history preserves the full exchange.

Real-World Scenario

A project manager uses Copilot to create a project status report.

The conversation includes:

Milestone summaries
Risk analysis
Resource concerns
Action items

Two weeks later, the manager needs to update the report.

Instead of creating a new conversation, they locate the previous conversation, review the earlier analysis, and continue working from that point.

This improves efficiency and preserves continuity.

Common Exam Misconceptions

Misconception 1: Previous conversations guarantee accurate information.

Reality:

Outputs should still be reviewed and verified.

Misconception 2: Conversation history bypasses permissions.

Reality:

Security and access controls remain enforced.

Misconception 3: Previous conversations are only useful for viewing old responses.

Reality:

They can also be continued, updated, and expanded.

Misconception 4: Saved prompts and conversation history are the same thing.

Reality:

Saved prompts store reusable instructions, while conversation history stores entire interactions.

Best Practices for Managing Conversation History

Use clear and descriptive conversation topics.
Revisit successful conversations when appropriate.
Reuse effective prompts.
Review previous outputs before acting on them.
Verify information before making decisions.
Protect confidential information.
Follow organizational governance policies.
Continue conversations when additional context is helpful.

Key Exam Takeaways

For the AB-730 exam, remember:

Previous conversations store past interactions between users and Copilot.
Conversation history helps users continue work without starting over.
Users can revisit prompts, outputs, and discussions.
Previous conversations improve productivity and context retention.
Conversation history can support verification and auditing.
Security permissions continue to apply.
Conversation history does not grant additional access rights.
Saved prompts and conversation history are different concepts.
Users should review and verify AI-generated outputs.
Previous conversations help preserve knowledge and support ongoing work.

B, C, and D reduce the usefulness of conversation history and make information harder to locate.

Go to the AB-730 Exam Prep Hub main page

AB-730, AI, Artificial Intelligence (AI), Microsoft Certification June 6, 2026

Save a prompt (AB-730 Exam Prep)

This post is a part of the AB-730: AI Business Professional Exam Prep Hub.
This topic falls under these sections:
Manage prompts and conversations by using AI (35–40%)
   --> Create and manage prompts in Microsoft 365 Copilot
      --> Save a prompt

Note that there are 10 practice questions (with answers) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

As users become more experienced with Microsoft 365 Copilot, they often discover that certain prompts consistently produce high-quality results. Rather than recreating these prompts each time, users can save prompts for future use. Saving prompts improves efficiency, promotes consistency, and helps users build a personal library of effective AI instructions.

For the AB-730: AI Business Professional exam, it is important to understand the purpose and benefits of saving prompts, when saved prompts should be used, and how prompt reuse can support productivity across business workflows.

Saving a prompt does not change how Copilot generates responses. Instead, it provides a convenient way to store and reuse effective prompt instructions that have proven useful for recurring tasks.

What Is a Saved Prompt?

A saved prompt is a prompt that a user stores for future reuse.

Instead of repeatedly typing the same instructions, users can:

Save the prompt.
Retrieve it later.
Modify it as needed.
Reuse it for similar tasks.

Saved prompts help standardize common business activities and reduce repetitive work.

Why Save a Prompt?

Many business tasks occur repeatedly.

Examples include:

Creating weekly status reports
Summarizing meetings
Drafting customer communications
Generating project updates
Analyzing sales performance
Preparing executive briefings

If a prompt consistently produces useful results, saving it can improve efficiency.

Benefits of Saving Prompts

Increased Productivity

Users do not need to recreate complex prompts each time.

Instead of writing:

Create a one-page executive summary highlighting risks, milestones, budget status, and next steps.

every week, the prompt can be saved and reused.

This reduces effort and saves time.

Consistency

Saved prompts help produce consistent outputs.

For example:

A manager may want all project updates to follow the same structure:

Executive summary
Milestones
Risks
Budget status
Action items

Using the same saved prompt helps maintain consistency across reports.

Reduced Errors

Recreating prompts manually may lead to:

Missing instructions
Inconsistent wording
Forgotten requirements

Saved prompts reduce the likelihood of accidentally omitting important guidance.

Improved Prompt Quality

Over time, users often refine prompts through experimentation.

Once a prompt consistently produces high-quality results, saving it preserves that work for future use.

Common Business Use Cases for Saved Prompts

Meeting Summaries

Example prompt:

Summarize this meeting for executives. Include decisions, risks, action items, and upcoming deadlines.

A user may save this prompt because it is used frequently.

Executive Briefings

Example prompt:

Create a one-page executive briefing focused on business impact, risks, opportunities, and recommended actions.

This prompt can be reused across multiple projects.

Customer Communications

Example prompt:

Draft a professional customer response that is concise, empathetic, and action-oriented.

Customer service teams may use this repeatedly.

Data Analysis

Example prompt:

Analyze the data and identify trends, anomalies, business risks, and recommendations.

This can support recurring reporting activities.

When Should You Save a Prompt?

Prompts are good candidates for saving when they are:

Frequently used
Well tested
Consistently effective
Applicable to recurring tasks

Good Candidates for Saved Prompts

Weekly reports
Monthly summaries
Project updates
Meeting recap requests
Customer service templates
Executive communications

Poor Candidates for Saved Prompts

Highly unique or one-time requests may not provide enough future value to justify saving.

Example:

Analyze the impact of a specific event that occurred yesterday.

The prompt may never be used again.

Creating Effective Prompts Before Saving Them

A prompt should ideally be refined before it is saved.

Users often follow a process such as:

Step 1

Create an initial prompt.

Step 2

Review the response.

Step 3

Adjust the wording.

Step 4

Test again.

Step 5

Save the prompt once it consistently produces desired results.

This process helps ensure the saved version is effective.

Saved Prompts and Reusability

The most valuable saved prompts are often reusable across multiple situations.

Less Reusable

Summarize the March 14 budget meeting.

More Reusable

Summarize this meeting and identify key decisions, risks, and action items.

The second prompt can be used repeatedly with different meetings.

Customizing Saved Prompts

Saved prompts are not necessarily fixed.

Users can:

Modify details
Change audiences
Add context
Adjust output formats

The saved prompt serves as a starting point.

Example

Saved prompt:

Create an executive summary of this project.

Modified version:

Create an executive summary of this project for senior leadership and include financial impacts and major risks.

The saved prompt accelerates the process while allowing flexibility.

Organizing Saved Prompts

As users build prompt libraries, organization becomes important.

Common categories include:

Meetings
Communications
Reporting
Data analysis
Project management
Customer service

Organized prompt collections help users quickly locate useful prompts.

Prompt Templates vs. Saved Prompts

These concepts are related but not identical.

Prompt Template

A reusable structure that contains placeholders.

Example:

Draft an email to [Audience] regarding [Topic].

Saved Prompt

A stored prompt ready for reuse.

Example:

Draft a professional email to customers announcing a planned service interruption.

Both concepts support efficiency and consistency.

Sharing Saved Prompts

Organizations may develop prompt libraries that employees can reuse.

Benefits include:

Standardized communication
Consistent reporting
Reduced learning curves
Improved prompt quality

Shared prompt collections can help teams adopt AI more effectively.

Responsible AI Considerations

Saving a prompt does not eliminate the need for:

Human review
Fact-checking
Verification
Compliance checks

Users should still:

Review outputs
Validate information
Follow organizational policies

A saved prompt can improve efficiency, but responsible oversight remains necessary.

Real-World Scenario

A project manager creates a prompt that generates excellent weekly status reports:

Create a one-page project update including milestones, risks, budget status, and next steps.

After refining and testing it over several weeks, the manager saves the prompt.

Each week, the manager can reuse the prompt with updated project information rather than creating new instructions from scratch.

This improves consistency and saves time.

Common Exam Misconceptions

Misconception 1: Saving a prompt guarantees accurate responses.

Reality:

Outputs should still be reviewed and verified.

Misconception 2: Saved prompts cannot be modified.

Reality:

Saved prompts can often be adjusted to fit specific situations.

Misconception 3: Only long prompts should be saved.

Reality:

Any frequently used and effective prompt may be worth saving.

Misconception 4: Saved prompts replace human judgment.

Reality:

Users remain responsible for reviewing and validating outputs.

Best Practices for Saving Prompts

Save prompts that are used frequently.
Refine prompts before saving them.
Organize prompts by task or business function.
Use clear and descriptive names.
Update prompts when business requirements change.
Continue reviewing AI-generated outputs.
Share useful prompts when appropriate.
Focus on reusable prompt structures.

Key Exam Takeaways

For the AB-730 exam, remember:

A saved prompt is a reusable prompt stored for future use.
Saving prompts improves productivity and consistency.
Frequently used prompts are good candidates for saving.
Saved prompts reduce repetitive work.
Effective prompts should typically be refined before being saved.
Saved prompts can often be modified and customized.
Prompt libraries can support team-wide AI adoption.
Saved prompts do not bypass the need for verification.
Human review remains important.
Saving prompts is a practical way to manage recurring AI-assisted tasks.

A, B, and D encourage inappropriate reliance on AI-generated content.

Go to the AB-730 Exam Prep Hub main page