Manage, monitor, and secure AI systems
18 practice questions under this official exam objective (AI-103), each with the answer key, an option-by-option explanation and instant feedback.
Questions under this objective
18 total · page 1 of 1Mitigating Prompt Injection in Multimodal AI with Foundry
This question tests the application of Azure AI Content Safety features to mitigate indirect prompt injection attacks in multimodal generative models.
Assessing Agent Responses with Foundry Evaluators
This question tests the distinction between RAG evaluators and groundedness evaluators in Microsoft Foundry for assessing AI response accuracy. The co
Improving Agent Response Completeness in Microsoft Foundry
This question evaluates the effectiveness of using evaluation flows to block incomplete AI responses. The correct approach involves understanding that
Restricting Azure AI Language Access with Virtual Network Rules
To limit Azure AI Language in Foundry Tools to specific Azure processes with minimal setup, use virtual network rules in the resource's built-in firew
Rotate a compromised Azure AI Search query key without downtime
This AI-103 question tests zero-downtime rotation of a suspected-compromised Azure AI Search query key: create a new query key, repoint the web app to
Return self-harm severity with Content Safety in Python
This AI-103 hotspot tests the Azure AI Content Safety Python workflow for moderating a comment string. The correct flow is AnalyzeTextOptions(text=com
Securing Azure AI Search with Private Endpoints and RBAC
This question tests securing Azure AI Search by combining network isolation via Private Link with granular access control using Azure RBAC roles.
How Should Apps Access Foundry and Select a Deployment?
This AI-103 hotspot asks how three REST API apps should authenticate to a Microsoft Foundry resource and route to separate GPT-3.5 deployments. The pa
Which control protects Agent1 from hidden product-sheet instructions?
Prompt Shields in Microsoft Foundry content safety is the control that addresses indirect prompt injection from product-sheet content. The answer is A
Diagnose Foundry Agent Issues: Model Availability, Limits, and Inference
Learn to map Microsoft Foundry agent errors like HTTP 429 and 'No relevant information found' to the correct diagnostic logs. This page explains how t
Implementing Document-Level Security Filtering in Azure AI Search
Azure AI Search document-level security is implemented by tagging each indexed document with allowed security groups, resolving the requesting user's
Does Content Safety Monitor Online Activity Optimize Filters?
Microsoft Foundry content filter tuning requires testing sample questions against moderation thresholds, not observing production traffic. The Content
Moderate Input Text Against a Custom Blocklist in Foundry
This AI-103 hotspot tests Content Safety custom blocklist moderation from Python. The answer is an AnalyzeTextOptions request with text, blocklist_nam
Azure AI Agent Guardrails: Tool Call Intervention for MCP
To prevent an Azure AI Agent from sending violence-tagged content to an MCP server, you must use the tool call intervention point for guardrails, whic
Optimizing Content Filter Configurations in Azure AI Foundry
This question addresses how to optimize content filter configurations in Microsoft Foundry using Azure AI Content Safety. The Moderate text content fe
Does Protected Material Detection Optimize Content Filter Configuration in Foundry?
Optimizing an Azure Microsoft Foundry agent's content filter requires testing sample questions against severity thresholds and blocklists, not running
Does the Safety Metaprompt Test Content Filter Configurations?
This AI-103 scenario asks whether the Safety metaprompt in Content Safety Studio can optimize a Foundry content filter by running sample questions aga
How Do You Restrict Azure Language Resource ta1 to vnet1?
Azure Language in Foundry Tools restricts inbound access by configuring the resource's own virtual network settings, not by filtering vnet1 traffic wi