Microsoft Copilot Agent Functionality

What is the fundamental characteristic of an agent in Microsoft Copilot 365?

https://www.youtube.com/watch?v=K3lyfJ-2RXQ

• A multi-tasking general-purpose AI.

• An AI that requires advanced coding and AI degrees to build.

• A custom version of Microsoft Copilot designed to accomplish one specific task.

• An agent that automatically learns and adapts to all user needs without specific training.

Answer explanation

The correct choice highlights that Microsoft Copilot 365 can be customized for specific tasks, making it a tailored solution rather than a general-purpose AI or one that requires advanced skills.

When building a new agent in Microsoft Copilot Studio, what is the recommended initial approach for new users?

• Directly using the 'Configure' tab for detailed setup.

• Starting with the 'Describe' tab and following the guided instructions.

• Uploading pre-built agent templates from an external source.

• Writing advanced AI code to define the agent's behavior.

Answer explanation

Starting with the 'Describe' tab provides a guided approach for new users, making it easier to understand the setup process and ensuring they don't miss important configuration steps.

For effective email management, especially when dealing with diverse email types, what is the recommended strategy for building agents?

• Develop a single, highly complex agent capable of handling all email types.

• Create a separate, specialized agent for each distinct type of email (e.g., customer orders, vendor refunds).

• Rely solely on generic email templates provided by the Copilot system.

• Manually categorize and pre-process all incoming emails before agent interaction.

Answer explanation

Creating separate, specialized agents for each email type allows for tailored responses and efficient handling of diverse inquiries, improving overall email management effectiveness.

Why is it beneficial to train an email-answering agent with examples of your own written emails?

• To increase the agent's processing speed for email responses.

• To enable the agent to understand and replicate your unique tone and writing patterns.

• To allow the agent to access and synthesize information from public websites.

• To reduce the need for any manual input or oversight in the future.

Answer explanation

Training the email-answering agent with your own emails helps it understand your specific tone and writing style, allowing for more personalized and accurate responses that reflect your unique communication preferences.

What was the primary criticism leveled against the initial set of instructions provided for the AI agent?

• They were overly prescriptive, limiting the agent's adaptability.

• They lacked sufficient detail and specificity for effective agent performance.

• They contained contradictory directives, causing operational ambiguity.

• They were too brief, failing to cover all necessary operational scenarios.

Answer explanation

The primary criticism was that the initial instructions lacked sufficient detail and specificity, which hindered the AI agent's ability to perform effectively in various scenarios.

Which structured prompting methodology was employed to generate more comprehensive instructions for the Copilot agent?

• Situation, Task, Action, Result (STAR)

• Goal, Context, Expectation, Source (GCES)

• Problem, Solution, Benefit (PSB)

• Audience, Purpose, Tone, Structure (APTS)

Answer explanation

The GCES methodology focuses on defining the Goal, Context, Expectation, and Source, which helps in creating detailed instructions for the Copilot agent, making it the correct choice over the other options.

To enable the AI agent to learn a specific tone and voice, what type of data was uploaded into its "Knowledge" section?

• Proprietary software code

• Historical financial reports

• Past email communications

• Customer service call recordings

Answer explanation

Past email communications provide a rich source of conversational tone and style, making them ideal for training the AI agent to learn a specific voice. Other options do not focus on tone as effectively.

What is the primary function of the "Code interpreter" capability in Copilot Studio?

• Generate visual aids like images and art.

• Convert natural language into code, create visualizations, solve math problems, and analyze data.

• Auto-detect the tone of incoming emails.

• Integrate with external calendars for real-time availability.

Answer explanation

The primary function of the 'Code interpreter' in Copilot Studio is to convert natural language into code, create visualizations, solve math problems, and analyze data, making it a versatile tool for developers.

What is the maximum number of suggested prompts that can be added to an agent in Copilot Studio?

• Three

• Four

• Six

• Unlimited

Answer explanation

The maximum number of suggested prompts that can be added to an agent in Copilot Studio is six. This allows for a diverse range of inputs to enhance the agent's performance.

What is a key advantage of using a custom-built agent in Copilot Studio compared to a general Copilot chatbox for specific tasks?

• Agents can only process text-based inputs, while general Copilot can handle images.

• Agents provide generic responses, whereas general Copilot offers tailored solutions.

• Agents leverage specific instructions and custom knowledge for more relevant and tailored responses.

• General Copilot can only be used by one person, while agents can be shared across an organization.

Answer explanation

The correct choice highlights that custom-built agents utilize specific instructions and tailored knowledge, enabling them to provide more relevant and personalized responses for specific tasks compared to a general Copilot chatbox.