Embedded Files
Kore.ai is a no-code/low code platform with a powerful user interface that allows you to build, deploy and manage chat bots. It uses an award-winning NLP or Natural Language Processing based on three engines - Machine Learning, Fundamental Meaning, and Knowledge Graph.
A Design Brief helps people organize the purpose of their creation and their plans for bringing their idea to fruition. This keeps them from going willy-nilly in all different directions.
In this case, the question is: What kind of service will your bot provide and to whom? Your Design Brief needs to provide details about the type of content you will build within your bot.
Before diving in, it will be valuable for you to become familiar with a few terms that are used when building chatbots. It can be easy to misinterpret the meaning!
Virtual Assistant/bot/chat bot - All of these terms refer to the chatbot you are building at Kore.ai.
Utterance - This is what the user has said. A chatbot user might say, Get me an agent. Check my balance. How much do I owe? Those are all examples of an utterance. It is anything a user types into the chat window.
Intent/Use Case - This is what the user actually wants. For example, the utterance of Get me an agent means the user has the intent or desire to be transferred to a human being.
Dialog Tasks - Bot conversations that result in something being accomplished on the user's behalf.
Node - A section or piece of a conversation.
Entity - A piece of data you collect from a person, like a name, date or place.
Context - This is data that is saved into the bot's memory. Some data is only saved for a short time and some data is saved for a really long time.
Knowledge Graph - Simple Q&A or FAQs designed to provide information only.
Talk to Bot - This is a little feature that allows bot builders to test if their conversation designs are working properly.
Channel - This is where the bot is deployed/hosted. Examples could be Twitter, Teams, Google Chat, Messenger, the Web, etc.
Deploy - This means to send off for the first time. The phrase deploy your chatbot means you are setting it up for the first time on a specific channel. You will only need to deploy once per channel.
Publish - This will send your new updates to your deployed channels.
Video transcript
Video transcript
Chatbots can be a great way to introduce kids to coding. A design brief would be an excellent organizer for helping them plan what their bot can do. Let's take a look at a bot I've created. Let me start by saying - Don't be afraid. This is not made for kids. It's made for business so there are a lot of technical features but very little will apply to a personal user. The basic things you will want to learn about are the main conversation pieces: Dialog Tasks and Knowledge AI, and how to Deploy your bot on the web and publish your updates.
The knowledge graph is a simple question-and-answer type of dialog. You add a question and answer, train the bot and then check it with Talk to Bot.
The Dialog Tasks are more complex and usually result in the bot performing some action such as an API call, a transfer or data collection. You build these by adding nodes to the conversation. Each node has a panel that allows you to further configure the option. In this task, we are making a change to a data table so we collect the users email address and new address, make the update and then conclude the conversation. Dialog Tasks are immediately available to Talk to Bot so you don't have train the bot in order to test it out.
When you open Talk to Bot, you can also open the Debug log. This shows you how each node is processed and really helps you find errors. That doesn't seem important now, but it will later on as you learn more and have to start troubleshooting problems. Over here, we can see what variables exist for this conversation. Let me show you how this Dialog Task works using Talk to Bot. You can see the log here and then over here under Session Context and Variables, you can see the variables. Sound confusing? Sound like something you could never learn? Maybe - but believe me, you can learn. Just give it time.
Now let me show you how to Deploy your bot to the web. Go to Deploy, Channels and choose the Web/Mobile Client. Here you can get a URL and you can design your chat window. Here's what mine looks like. This only reflects the most recent publication of my bot. As I add more content, I will need to republish in order for this chat window to be updated.
If you share this link with others and they spend time chatting with your bot, you can actually see the stats under the Analyze tab. This can also help you to identify common errors and problems with your bot.
This course has been curated by Mind Garden Media, LLC.
Click the icons to visit our socials (and follow us) on Facebook, Twitter (X) and YouTube!
Page updated
Google Sites
Report abuse