AB-100: Agentic AI Business Solutions Architect Exam Prep

Outline the AB 100 exam syllabus across plan, design, and deploy AI powered business solutions, covering ROI, build/buy/extend decisions, and use of Azure AI, Copilot, Foundry, and Dynamics 365.

Explore the evolution from artificial intelligence to generative ai and the role of large language models and transformers. Discover how foundation models and fine-tuned apps empower Copilot and ChatGPT.

Master prompt engineering for generative AI applications like Copilot, ChatGPT, and Gemini by crafting precise prompts with clear goals, context, expectations, and source.

Explore how Microsoft 365 Copilot uses Microsoft Graph grounding and retrieval augmented generation to access tenant data securely and deliver context-aware answers.

Understand the Microsoft 365 Copilot architecture from user prompt to summarized response, including Graph API context, knowledge and action plugins, and Azure OpenAI foundation models within your tenant.

Get started with the Copilot work and web experiences in Microsoft 365. Create simple agents, explore pre-built options, and connect to knowledge sources like Teams, SharePoint, Outlook, and web URLs.

Explore managing copilot adoption, licensing, and pay as you go billing in the Microsoft 365 admin center, with connectors, usage insights, budget alerts, and Azure subscription integration.

Explore the principles of responsible AI with Microsoft Copilot, including fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability, plus how to avoid overreliance and fabrication through citations.

Demonstrates a hands-on lab use case of the copilot analyst agent in m365, analyzing an uploaded data file with Python code, and producing charts and key insights.

Draft a structured Word document with Copilot in Word, creating a generative AI intern job opening, including stipend, application details, required skills, and company information.

Copilot visualizes text from a Word document into tables, turning ingredients with amounts into an ingredients table and directions into steps, in a hands-on lab showcasing tabular visualization.

Conduct a hands-on lab using Copilot in Excel online to format the data as a table and generate insights, pivot charts, and outlier detection by date and region.

Use Copilot to draft an end-to-end PowerPoint presentation from scratch on fruit juices, including templates and designer options.

Explore Copilot in Outlook to draft emails, reply professionally, and coach your writing. Generate concise email summaries and improve tone and detail with coaching by Copilot.

Summarize chat threads with Copilot for Teams to extract key takeaways and citations from a Gateway Corporation–Future Works Corporation sales chat. Track budget changes and a Zoom call request efficiently.

Learn to create a custom ESG framework education agent in Copilot for M365 using the describe tab and natural language input, delivering clear bullet-point explanations.

Leverage the copilot agent designer’s configure tab to build a Carbon Ops ESG agent from grounding documents and PDFs, enabling code interpreter and image generation tailored to grounded knowledge.

Discover how Microsoft Purview delivers information protection and governance across the Microsoft 365 Copilot AI ecosystem, enabling data discovery, sensitivity labeling, data loss prevention, and multi-cloud compliance.

Learn information protection with Microsoft Purview by applying sensitivity labels and access controls to documents across finance and HR teams for Future Works Corporation, including Copilot for M365.

Learn to implement information protection in Microsoft Purview. Build sensitivity labels, auto-labeling policies, and classifiers to protect GST numbers and other sensitive data across SharePoint, OneDrive, and M365.

Learn how Microsoft Purview communication compliance governs Copilot for M365 chats, detects inappropriate content and sensitive data, and enables templates, remediation workflows, and insider risk alerts.

Demonstrates creating and customizing communication compliance policies in purview, using pre-built and custom templates to detect sensitive info types like GST numbers, set locations, directions, and reviewers, and manage violations.

Explore how DSP for AI in Microsoft Purview helps govern AI workloads with policies and recommendations, including data discovery, DLP, and governance of Copilot for M365 and Copilot Studio agents.

Learn how compound AI systems and AI agents orchestrate tasks using LLMs, planning, and tools, with Copilot as the UI and Copilot Studio or Azure AI Foundry for implementation.

Explore how Microsoft Copilot Studio lets citizen developers build and extend agents with no-code to low-code tooling, connect to 1200+ data connectors, and integrate with Foundry for scalable AI workflows.

Explore Copilot Studio licensing options, including pay-as-you-go and a 25,000-credit monthly commitment, pricing, free trials, and how credits are consumed by generated answers and agent actions.

Extend Microsoft 365 Copilot with Copilot Studio or Azure AI Foundry to build domain-specific agents that solve department-level business problems, maximizing ROI on licenses and moving toward department oriented intelligence.

Explore the three types of agents in Microsoft Copilot Studio—prompt and response agents, task agents, and autonomous agents—and how they use knowledge bases, dataverse workflows, and multi-agent workflows.

Explore the generative AI orchestrator and agent planning in Microsoft Copilot Studio, detailing dynamic runtime decisions that order tool and plugin execution for user intents.

Create a web grounded standalone agent in Microsoft Copilot Studio by linking an external website as knowledge source, configuring prompts and triggers, then publish to Teams and Microsoft 365 Copilot.

Build an embedded SharePoint agent for Microsoft 365 Copilot that uses SharePoint PDFs and a Word document as a knowledge base, with on behalf of authorization flow and published citations.

Explore embedded versus standalone agents in Microsoft Copilot Studio, delivering in-chat contextual assistance across 365 apps, or a separate tab with enterprise knowledge beyond 365.

Add a prebuilt MSN weather data connector to your agent to retrieve current weather for a location, and use the generative AI orchestration planner to generate a concise report.

Create and attach a custom prompt action as a tool to a Copilot Studio agent to perform HR resume screening, producing structured JSON output from resumes.

Design, test, and deploy end-to-end agent flows in Copilot Studio by integrating a resume screening custom prompt action with Power Automate, and auto-email results via Office 365.

Demonstrates building a sentiment analysis flow with AI Builder in the Power Platform Copilot Studio, classifying reviews as positive, neutral, or negative with confidence, and routing results via email.

Integrate an MCP server with a Copilot Studio agent and generate a three-month AI 102 study plan using Microsoft Learn modules.

Analyze the analytics tab in Microsoft Copilot Studio to track conversation sessions, engagement, customer satisfaction score, and usage metrics, then assess savings and ROI from tool executions.

Explore Copilot Studio's agent settings tab in the hands-on lab to configure security, channels, connections, and languages, and control generative AI orchestration, content moderation, and voice capabilities.

Explore how Microsoft Foundry orchestrates multi-agent ai workflows in Azure, with a model catalog, agent service, Foundry IQ, tools, governance, and M365 integration to scale safe, observable ai production.

Launch a Microsoft Foundry resource in Azure Portal, create a standalone project, and explore Foundry Studio to deploy models, monitor agents, and access the model catalog and guardrails.

Deploy a GPT 5.2 chat model from the model catalog in Microsoft Foundry via serverless API inferencing, then compare models with the leaderboard and trade-off charts for quality and cost.

Analyze queries to determine reasoning depth and dynamically route to the optimal model router, balancing cost and accuracy for enterprise workflows.

Learn to create and version a first agent in Microsoft Foundry, use multiple tools including web search and code interpreter, and review traces, memory, and version history.

Attach an MCP server to a Microsoft Foundry agent, register it as an external tool, and use MCP tools to search Microsoft Learn content and code samples.

Explore predictive AI services in Microsoft Foundry, including Azure speech, translation, and PII redaction, managed via a single AI services resource with endpoints and API keys.

Explore building scalable multi-agent workflows with Microsoft Foundry, Microsoft Agent Framework, and Dev UI, covering sequential and parallel patterns, memory stores, and human-in-the-loop integration.

Explore agent evaluation and red teaming in Microsoft Foundry, where blue, purple, and red teams test security through direct adversarial probing and automated risk testing with pyRIT.

Engage in a hands-on lab to evaluate agents in Microsoft Foundry using groundedness, retrieval, coherence, and accuracy evaluators, then perform red teaming with risk categories and attack strategies.

Configure guardrails and content safety filters in microsoft foundry to protect agents and foundational llms from jailbreaks, harmful content, and sensitive data leakage.

Learn to set up an AI gateway with Azure API Management inside Microsoft Foundry, including token management and rate-limiting governance. See how to onboard MCP servers and apply inbound policies.

Explore how the Microsoft Power Platform enables no-code to pro-code development for business users, integrating Power Apps, Power Automate, Power BI, Power Pages, Copilot Studio, and Dataverse for AI-driven solutions.

Explore canvas apps, model-driven apps, and power pages, their use cases, differences, and licensing, including user interface control, data handling in dataverse, and external facing websites.

Explore how Microsoft Dataverse serves as a single source of truth with relational tables and built-in security, enabling Power Apps, Power Pages, and Copilot workflows.

Explore how Power Platform environments and solutions enable end-to-end application lifecycle management, from development to production, using DevOps pipelines, unmanaged vs managed solutions, and region-aware deployments.

With Copilot in the make portal, create a custom reviews table in Dataverse for sentiment analysis, set columns and data types, and configure forms and views for a model-driven app.

Build a copilot-powered canvas app on a dataverse reviews table to capture reviews, run sentiment analysis, and trigger a power automate flow across two screens.

Create a sentiment analysis workflow with AI Builder in a Power Automate flow triggered by a Power Apps button, and store results in the Dataverse reviews table.

Integrate a Copilot studio sentiment analysis agent with a canvas app in Power Apps to perform real-time sentiment analysis on reviews and trigger a backend workflow.

Learn to convert an unmanaged solution into a managed one by bundling canvas apps, Copilot studio agents, AI workflows, and a Dataverse table for deployment across environments.

Explore Microsoft Dynamics 365, an AI-powered CRM and ERP suite that covers sales, service, and operations. See how Copilot and Dataverse empower pre-built and custom agents on the Power Platform.

Explore sales fundamentals and the end-to-end sales lifecycle, from identifying leads and qualifying opportunities to quoting, invoicing, and closing deals, with Dynamics 365 for sales automation and AI agents.

Explore dynamics 365 sales co-pilot’s prebuilt and custom agents, including the sales qualification agent and deals close agent, that summarize leads and opportunities and surface insights for smarter engagements.

Explore the Dynamics 365 sales portal within PowerApps, navigate accounts, contacts, leads, opportunities, and Copilot-powered summaries, while configuring sales qualification and deal-closure agents via the AI hub.

Dynamics 365 customer service streamlines case management with knowledge articles, omnichannel channels, and SLA tiers, empowering agents with Copilot to summarize issues, propose solutions, and draft responses.

Explore the Dynamics 365 customer service hub to manage accounts, contacts, cases, and queues, using Copilot, knowledge articles, and email templates to accelerate case resolution and predictive maintenance.

Discover Dynamics 365 Customer Insights, unifying data into real-time profiles and enabling targeted journeys with Copilot for personalized campaigns and segments across email, SMS, and push.

Explore a hands-on walkthrough of Dynamics 365 Customer Insights, creating journeys, segments, and triggers in PowerApps, and using Copilot to draft emails and automate customer engagements across channels.

Explore Dynamics 365 field service and Copilot AI use cases for mobile workforce management at customer sites, from case creation to on-site service and invoicing.

Explore Dynamics 365 Finance and Operations use cases, including the account reconciliation agent that automates bank reconciliation, and the collections agent that summarizes receivables and drafts proactive collection emails.

Explore Dynamics 365 supply chain management features, including product catalog, asset and cost management, procurement, and IoT-driven planning, with Copilot use cases for real-time warehouse workload and operational insights.