Foundry, Agent Extensibility, MCP and Computer Use
This topic tests custom models in Microsoft Foundry, Microsoft 365 agents, Copilot Studio extensibility, Model Context Protocol, Computer Use, reasoning, voice mode, Teams, and SharePoint optimization.
How to study for Microsoft AB-100
Treat each item as an architecture decision: identify the business process, select the Microsoft AI pattern, then add ALM, telemetry, security, responsible AI, and governance controls.
Core concepts
Concept 1
Microsoft Foundry custom models fit scenarios that need specialized model behavior, evaluation, deployment, and lifecycle control.
Exam cue: Use MCP when an agent needs standardized access to external tools or context providers.
Concept 2
Agent extensibility uses Copilot Studio, Microsoft 365, MCP, connectors, tools, and actions to reach systems and workflows outside the base agent.
Exam cue: Use Microsoft 365 agents when Teams, SharePoint, and Microsoft 365 work context are central.
Concept 3
Computer Use and agent behaviors such as reasoning and voice mode should be designed around risk, auditability, user experience, and process fit.
Exam cue: Use custom Foundry models when out-of-box model behavior cannot meet the requirement.
Risk pitfalls and guardrails
Choosing extensibility before defining action permissions and audit needs.
Guardrail: Avoid defaulting to custom agents, skipping grounding-data checks, or treating prompts and connectors as outside the release process.
Using Computer Use where a stable API or connector is safer and more reliable.
Guardrail: Avoid defaulting to custom agents, skipping grounding-data checks, or treating prompts and connectors as outside the release process.
Deploying a custom model without a validation and lifecycle plan.
Guardrail: Avoid defaulting to custom agents, skipping grounding-data checks, or treating prompts and connectors as outside the release process.
Memory anchors
Microsoft Foundry
Microsoft Foundry supports building, evaluating, and managing AI models, agents, and tools.
Custom Model
A custom model is adapted or built to meet specialized business behavior or performance needs.
Agent Extensibility
Agent extensibility lets an agent use tools, connectors, actions, or protocols beyond built-in behavior.
Model Context Protocol
MCP standardizes how agents connect to external context, tools, and systems.
Computer Use
Computer Use lets an agent interact with apps or websites when API-based integration is not available.
Reasoning Behavior
Reasoning behavior controls how an agent analyzes tasks, plans steps, or uses tools.
Voice Mode
Voice mode adds spoken interaction patterns and accessibility considerations to an agent.
Microsoft 365 Agent
A Microsoft 365 agent operates in Microsoft 365 contexts such as Teams and SharePoint.
Checkpoint rule
Do the check-up only after you can summarize each concept in one sentence and identify one dangerous pitfall from memory.
Knowledge Check (after reading)
Short check-up to confirm understanding of this module.
Check-up Questions
A manufacturer needs a model specialized for visual inspection of a proprietary defect type that available models cannot detect reliably. Which approach should be evaluated?
A team selects a model from the Foundry catalog. What must happen before production deployment?
Answer all questions to submit.
Next step personalized recommendations
Continue learning
Move forward only after this module is stable.
What is Pass Harbor?
Completely free exam prep for 317 U.S. exams.
- Practice questions
- Flashcards
- Study guides
- Mock exams
- No registration
- No paywall
- Start instantly
“No more expensive exam prep. Quality study tools should be accessible to everyone.”