Bash, the Crucial Exams Chat Bot
AI Bot
Copilot Core Concepts (AB-900) Flashcards
Microsoft 365 Certified: Copilot and Agent Administration Fundamentals AB-900 Flashcards
Study our Copilot Core Concepts (AB-900) flashcards for the Microsoft 365 Certified: Copilot and Agent Administration Fundamentals AB-900 exam with 35+ flashcards. Review key concepts as flashcards, a searchable table, or an interactive matching game to reinforce exam concepts.

| Front | Back |
| How can Copilot automate workflows? | By connecting across apps running predefined actions and generating scripts or formulas |
| How can Copilot help in Excel? | Analyze data create formulas build pivot tables and generate charts from plain language requests |
| How can Copilot help in Microsoft Word? | Draft and revise text generate summaries rewrite content and suggest tones and styles |
| How can Copilot help in Outlook? | Draft email replies summarize threads suggest meeting actions and prioritize messages |
| How can Copilot help in PowerPoint? | Create slide outlines design slide content suggest visuals and turn notes into professional slides |
| How can you ask Copilot to cite sources? | Request explicit citations ask for URLs or document references and ask for confidence levels |
| How can you reduce hallucinations? | Provide clear grounding use reliable sources include examples and request citations |
| How do tokens relate to cost and performance? | More tokens increase compute usage and may raise cost while enabling longer or more complex outputs |
| How does Copilot access Microsoft 365 data? | It uses Microsoft Graph and secure connectors to retrieve user files calendar emails and other content with consent |
| How does Copilot handle user privacy? | It follows organization policies restricts data exposure uses role based access and logs access for auditing |
| How should you evaluate Copilot outputs? | Check for factual accuracy relevance completeness tone and potential bias before using or sharing |
| What are best practices for crafting prompts? | Be specific provide context define desired format and give examples or constraints |
| What are safety filters in Copilot? | Mechanisms that detect and block unsafe or sensitive content and enforce compliance policies |
| What core functions does Copilot provide? | Drafting and editing text creating summaries generating ideas automating workflows and extracting insights |
| What does grounding mean? | Tying AI responses to external verified data such as documents databases or search results |
| What is a context window? | The amount of text the model can consider at once usually measured in tokens |
| What is a safety use case to avoid when using Copilot? | Asking Copilot to generate sensitive personal data or perform legal or medical advice without expert review |
| What is a system prompt? | A high level instruction that defines the AI role and behavior for the session |
| What is a user prompt? | The specific question or task the user asks Copilot to perform |
| What is chain of thought prompting? | Encouraging the model to reason step by step to improve complex problem solving |
| What is few shot prompting? | Providing a few examples in the prompt to teach the model the desired format or style |
| What is hallucination in AI outputs? | When the model generates incorrect or fabricated information presented as fact |
| What is iterative prompting? | Refining prompts and asking follow up questions to improve output quality |
| What is Microsoft Copilot? | An AI assistant that helps users in Microsoft 365 by generating content answering questions and automating tasks |
| What is multimodal capability in Copilot? | The ability to process and generate text along with other data types like images or tables |
| What is prompt engineering? | The practice of designing inputs to guide AI models toward desired outputs |
| What is prompt templating? | Creating reusable prompt patterns to standardize requests and outputs across users |
| What is retrieval augmented generation RAG? | Combining external documents retrieval with generation to produce grounded answers |
| What is temperature in AI models? | A parameter that controls output randomness with lower values producing more deterministic responses |
| What is the difference between a Copilot and an autonomous agent? | Copilot augments human work inside apps while an agent makes autonomous decisions and takes actions without continuous human control |
| What is the role of human oversight with Copilot? | Humans review validate and approve AI suggestions ensuring correctness and ethical use |
| What is zero shot prompting? | Asking the model to perform a task without providing examples |
| When should you provide examples in a prompt? | When you need consistent formatting complex output or a specific tone and style |
| When should you use RAG with Copilot? | For tasks that require factual accuracy or up to date information from your own data |
| Which Microsoft apps commonly integrate Copilot? | Word Excel PowerPoint Outlook Teams and OneDrive |