Yes, it’s easier than ever... you can now, with just a few simple steps, create your own AI agent to help you complete tasks and make the most of your time. 

First, let me define what an AI agent is. Think of it as a virtual assistant working on your behalf and, if needed, without any manual intervention. It takes over repetitive and boring tasks, making them a thing of the past. It organizes your workflows, enhances work quality, and reduces errors. It’s not just a simple function or sequence of actions. it’s an Interactive, Intelligent Virtual Assistant. 

Whether you are a teacher, lawyer, manager, accountant, doctor, or in any other profession, if you are using Microsoft Teams, the door is open to build and configure your virtual assistant without needing to code or possess advanced AI expertise.  Microsoft provides us with the ability to create our own AI Copilot declarative agents, ranging from code-free to pro-code options.

Additionally, you can share your agent with others, making it highly beneficial and a valuable addition to your team or company. 

(A declarative agent relies on the robust and scalable infrastructure of Microsoft 365 Copilot, tailored to address a specific aspect of your requirements.)

In this article, we will create a new agent from scratch to answer employees' questions about HR regulations or internal company policies. 

 Let’s build your first agent using Microsoft Teams by following these steps:

1.      In Microsoft Teams, click on the Copilot button located in the left pane. 

(To access the Copilot feature in Microsoft Teams, you'll need a subscription to Microsoft 365 Copilot.) 

2.     On the right, you will find a new panel dedicated to Copilot. Press “Create an Agent"

3.     You can choose one of the 3 available agent templates or configure your own. In this article we will configure the agent manually

· Name: Assign it a unique, understandable name and add a logo to represent it. 

· Description: Write a definition text to help users understand the agent’s purpose, especially if it is shared with others. 

· Instructions: Provide the agent with its "job description" by detailing what it should do, how it should function, and any restrictions or operational guidelines. This step is the most crucial, as the agent will always refer to these instructions before performing any action. Its responses will be refined based on the tasks provided. 

 For our current example, we will use the following instruction text: 

You are an AI assistant tasked with helping users find information related to HR company policies and regulations based on provided knowledge. You should only respond by referencing the company's pre-provided policies and never speculate or provide information beyond the given context.

- Respond to user inquiries clearly and concisely, ensuring you reference the relevant sections of provided HR documentation whenever applicable.

- If the user's query is outside the scope of the knowledge provided or unclear, ask clarifying questions or inform the user that the specified information is not available.

- Avoid making assumptions or providing advice that is not explicitly supported by company policies or regulations.

- Use professional and neutral language at all times.

4. Knowledge: Add knowledge sources, such as data or documents, to be used as references. Additionally, you can include URLs of public websites or ask the agent to search internally to gather information. (It is important to note that you can connect to SharePoint as a knowledge source, but this requires a Microsoft 365 Copilot license for SharePoint sites.)

5. Starter prompts: Add text, information, or questions that guide the Copilot agent, helping it focus on specific targets and initiating a conversation. (Creating a prompt In Copilot studio can sometimes be confusing and challenging. You can refer to Microsoft’s Prompts Gallery: https://copilot.cloud.microsoft/en-us/prompts/all to explore many examples of prompts, which you can use or adapt to your needs.) 

Starter Prompts examples of Titles and Messages:  

6.     Finally, create your agent and start using it or sharing it. Make sure to select the appropriate sharing options while considering the confidentiality and privacy of your knowledge

With these simple steps, you're ready to build your very own AI agent, enhance productivity, streamline your tasks, and transform your workday.

Start building your AI Agent in Microsoft Teams and meet your virtual Assistant!