Microsoft Copilot Quiz

What does the term "copilot" refer to in the Microsoft ecosystem?

• A physical robot assistant

• A cloud-based virtual assistant using an LLM

• A software program installed on a local machine

• A type of search engine

What is the foundation of the "Copilot stack"?

• Data security

• AI infrastructure and models

• User interface design

• Cloud computing

Which of the following is NOT a way to interact with copilots in the Microsoft Cloud?

• Adopt an existing copilot

• Extend a copilot with plugins

• Build a custom copilot

• Download a copilot to a local machine

What is the primary benefit of building a custom copilot?

• It's always cheaper than using existing copilots.

• It allows for personalized conversational AI experiences integrated with company data.

• It requires no coding skills.

• It always outperforms pre-built copilots.

What is Azure AI Studio?

• A platform for creating and deploying custom copilots.

• A suite of pre-built copilots.

• A data storage service.

• A plugin for Microsoft Teams.

What is a key feature of Copilot Studio?

• It requires advanced coding skills.

• It focuses on command-line interfaces.

• It's a SaaS-based graphical tool for building conversational AI.

• It only works with structured data.

What is the meaning of the expression "data is the fuel that powers AI"?

• AI models require large amounts of energy to run.

• Data quality significantly impacts the performance of AI models.

• AI models can only process numerical data.

• AI models learn best from unstructured data.

What does "groundedness" refer to in the context of LLM responses?

• The grammatical correctness of the response

• The length of the response

• Whether the response is rooted in reality or a specific context

• The speed of the response

What is Retrieval Augmented Generation (RAG)?

• A type of large language model

• An architectural design for including grounding data in an LLM prompt

• A data storage technique

• A method for training LLMs

What is the role of a search index in RAG?

• To store the entire training data of the LLM

• To organize content for efficient retrieval of relevant information

• To generate embeddings for the entire data source

• To evaluate the quality of LLM responses