Conversational AI Intro

Hello, everyone. Welcome to the SPOT solutions channel. My name is Nancy Amaro. I’m a project manager, business analyst, and power platform citizen developer. Today, I’m thrilled to introduce a miniseries focused on developing a conversational AI using Copilot Studio.

This series will take you on a journey from understanding the basics of AI to actually creating an intelligent and engaging AI assistant. So, whether you’re new to AI or looking to enhance your existing skills, this series is designed to guide you through the entire process. So, let’s get started. By the end of this mini series, you will be proficient in designing, building, and deploying conversational AI solutions using Copilot Studio. You will gain practical skills to create conversational AI, automate processes, and enhance business performance.

In this first video, I will introduce Copilot Studio by understanding the role of conversational AI in modern applications and getting familiar with how conversational topics work. Conversational AI plays a key role in modern applications, transforming the way we interact with technology. It leverages natural language processing to respond to human language in a valuable and contextually appropriate manner. From virtual assistants like Siri and Alexa to customer service agents, conversational AI is at the forefront of enhancing user experience. It enables 20 fourseven customer support, instant responses, and personalized interactions, improving customer satisfaction and operational efficiency.

With tools like Copilot, developing conversational AI has become more accessible. These tools empower citizen developers to build intelligent AI assistance that can integrate with SharePoint online, Teams, and leverage thousands of connectors, opening up a world of possibilities for innovative applications. In essence, conversational AI is not just a trend but a significant step to a more interactive and intuitive technological future. So let’s get started by downloading your free trial of, Copilot Studio. So head over to the official Microsoft page and click on try free.

Once you access your account, let’s go ahead and discover Copilot Studio features and capabilities. Once you have access to your account, you can create your 1st conversational AI by clicking on the, create. So, locate the create button using the left hand side navigation, and then click on new Copilot. This will display a form so that we can provide basic information about our Copilot. In this case, I need to provide a proper name.

So the name of this Copilot will be Spot Help Desk for SharePoint support. We need to provide a proper description. The Copilot helps to answer common questions related to SharePoint online usage. So that is my description. And in terms of the instruction, so the instruction, is helpful to tell the Copilot what its primary purpose is and how it should act.

In this case, I am indicating that your main goal is to provide clear and accurate responses to inquiries from Spot Solutions employees regarding the proper use of SharePoint Online. Please ensure your responses are concise and conveyed in a formal yet approachable manner. Then we have the option to identify the primary language. In this case, I will leave the default as English. And the last, part here is to identify knowledge sources.

Adding knowledge sources allows you to connect your conversational AI to internal resources, like a SharePoint site or a local file. You can also use external resources like a public website. We will dive into these, sources in later videos. So, once you have identified basic information for your Copilot, go ahead and click on create. That will set up your Copilot.

And once we have that ready, I will go ahead and highlight a few key features of your Copilot, studio. So, the first one is that we have the ability to customize our conversational AI. There are many different scenarios in which a conversational AI is helpful. You can create an internal help desk to assist employees with IT related questions. You can gather feedback from customers to improve your brand or products.

You can create an employee onboarding assistant that guides new employees to the process, and many more scenarios in which a conversational AI can be leveraged. Most organizations benefit from improved and efficient customer service while freeing up human agents for more complex tasks. So now that we have access to our Copilot studio and we have set up our first, conversational AI, the first thing that I will show you is how topics work. Copilot Studio provides a graphical low code interface for creating conversational AI models, making it accessible even to those without a background in data science or development. So let me show you how easy and simple it is to create a basic conversation.

So here on the screen, we see, 2 different types of, topics. We have custom topics and system topics. Every Copilot comes with a list of, predefined, topics and system topics. Both are editable. System topics are necessary to help your Copilot run effectively.

Without them, your Copilot may not work as intended. And one of the key important factors here to note is that we cannot delete these system topics, but we can always inactivate them, by clicking on the 3 dots and, identifying here the status toggle and turning that on and off. So the different, topics here can always be managed by enabling or disabling using the toggle option. Now, system topics are predefined in Copilot and they help the end users to get going with the conversation. So, these are basic and trivial ways of starting a conversation.

It means that we don’t have to create these topics from scratch. They’re here available for us to leverage. So conversations such as initiating the conversation or ending the conversation, maybe escalating the conversation to a live agent. And then another one is having a fallback or having multiple topics match. Maybe there’s an error on some of the, topics.

Or we might want to reset the conversation to the beginning. So, these system topics allow for your conversational AI to be optimized to leverage all of these native features within Copilot. Now, custom topics. This is what we used to customize, the different conversations that we will have with, our end users. Now, let me show you how simple it is to create a conversation.

Conversations are built around trigger phrases and messages. So, let me show you this simple topic so that you can see, how trigger phrases and messages work. So, we have a simple topic here and a list of trigger phrases. Our trigger phrases are when are you closed, when are you open, store hours, and daily open hours. So, these are the different ways in which a user might start a conversation, and those are the trigger phrases.

Once the Copilot identifies any of those trigger phrases, then it will respond to the end user with a message. In this case, based on the sample trigger phrases, the end user will receive the following message, I’m happy to help with store hours, followed by another message identifying these specific hours for certain locations. Now, in order for us to quickly test our Copilot, we can see that on the right hand side we have a test button. When you click there and use any of the trigger phrases, so, for instance, let’s go ahead and ask when are you closed? Then the Copilot will respond with the message being configured here.

So we can see I’m happy to help with store hours and then the information that came after, which is the specific location hours. Then we have the end of conversation topic. And this end of conversation is basically redirecting this conversation to that other topic, and that other topic handles the end of the conversation. So we have here, did that answer your question? So we can say yes or no.

If we say yes, then it will ask for a rating. We can go ahead and rate this experience, and then it will say, thank you for your feedback, can I help you with anything else? If we say no, then that will end the conversation. So, the main structure for designing the different conversations within our conversational AI works around identifying different triggers and system messages, that interact with the end user. And we can quickly test those trigger phrases by clicking on the test icon and using those sample phrases in order to test out, the different steps identified in our conversation.

In our next video, I will show you how to design your own conversation. So, that, that’s it for this first video. Thank you for watching.