Sinch Community

Entities are relevant pieces of information for your bot to extract from a user's expression and remember so that it can re-use it. Therefore, entities should only be used if these pieces of information are needed in the conversation.   For example, in the message "I want to order a large pepperoni pizza", entities like  "pizza type" and "pizza size" are relevant pieces of information to place a pizza order.   If an entity is successfully detected, the bot will store this entity as a variable.    Entities can be: Contextual entities Match entities System entities See how you can check if the user input matches your entity inside of a Collect input block.   See also: [New bot builder] Recognize expressions in intents [New bot builder] Add an intent to your bot [New bot builder] Train your bot   ... View more

Creating a different channel experience will be essential in providing a tailored user experience. It is important to consider the limitations of each channel when designing your messaging strategy. Here is a channel comparison to help you determine the most appropriate message types for each channel:   Message type WhatsApp Webchat Facebook Telegram Text message ✅ ✅ ✅ ✅ Buttons ✅max. 10 buttons ✅max. 10 buttons ✅max. 3 buttons ✅max. 3 buttons Carousel 🚫 🚫 ✅ 🚫 Media ✅ ✅ ✅ ✅ WhatsApp list  ✅ 🚫 🚫 🚫   Special restrictions on buttons About buttons, please bear in mind: URL and Call buttons do not work on WhatsApp. Instead, please use the Go-to buttons. WhatsApp and Facebook Messenger, are limited to 20 characters for buttons titles. Facebook and Telegram buttons can be max 3, whereas WhatsApp and Webchat can have up to 10 buttons. If you need to include more than 3 buttons for different channels, consider the following approaches: For WhatsApp, use WhatsApp Lists Messages to offer more than 3 options. For Facebook, use the Carousel which allows more than 3 options.  For Webchat, you can use buttons to provide options more than 3.  For Telegram, optimize options beyond 3 using intents and numbers.    See also: Create a channel confirguration [New bot builder] How to organize your flow [New bot builder] Make a channel router flow   ... View more

Within a conversation, speakers remember pieces of information to steer the conversation in relevant ways. When building your chatbot on Sinch Engage, there are 2 solutions for remembering information about your customer: customer properties, which are customer details like their name or email address. You can create more customer properties under the Customers tab. Customer properties are relevant for multiple conversations. variables, which are pieces of info relevant to a particular chatbot conversation. These cannot be used by an agent. For instance, a variable could be the type of pizza that the customer wants to order. There are two ways of collecting variables: with a Collect info block, or under button click. Variables and customer properties both have a name and a value. The name is used to refer to the variable or property. The value is what the variable or property represents.   Please note that entities that are successfully detected by your chatbot will be stored as variables.   See also: Create a customer property [New bot builder] Collect input from your user [New bot builder] Save a button click under a variable ... View more

If your chatbot lives on WhatsApp, you can benefit from the text formatting WhatsApp allows in its messages.    To format your messages for WhatsApp, open any Text block and use the WhatsApp formatting documentation  to guide you.   In this example, we have formatted the message within a Text message block as follows:     This is how the message is displayed on WhatsApp:     Not sure how to set up a channel for your bot? Learn how to configure channels in Sinch Engage.    ... View more

Your bot's introduction message is the first message that the customer will receive. It's part of the predefined blocks on your canvas.   To give the best first impression, this introduction should already include a lot of what the user needs to have a successful conversation with the bot. If you haven't yet learned the basics of conversation design, we suggest you start in the Conversation design intro.    In Engage, your welcome message is added by default as a starting point in the conversation. This means that it will be sent when the conversation starts.     Best practices for writing an introduction message Once your bot's definition has been outlined in the chatbot design workflow, make sure that a couple of aspects of this definition are displayed in the welcome message: the bot's personality and tone the scope of the conversation In the example below, on the right you can see a good welcome message because: The bot is presented as a bot. This is always good so that customers do not think they are speaking to a human. The bot shows some personality traits. In this example, the name, emojis, and empathy included in the message give the customer the impression that this bot is welcoming, warm, and will listen carefully to what the customer says. The bot is presented as an FAQ bot. The last sentence outlines the purpose of the bot: to answer the customer's questions on yoga studios.  On the left, you can see a poor welcome message that does not bring any added value to the conversation. This increases the chances of your customer asking something out of scope and then being blocked in the conversation because it wasn't understood by the bot. Eventually, this leads to a frustrated customer who does not want to use the bot again.   To make your welcome message even clearer, you can also use reply buttons so that the topics covered by the bot are well defined. All you need to do is add a Buttons step under the Introduction message block on your bot canvas. The image below uses the same example as above but adding buttons.       See also: [New bot builder] Conversation design intro [New bot builder] Chatbot workflow [New bot builder] Canvas overview ... View more

Designing the way a chatbot converses is called conversation design. It's a combination of technology, psychology, and language that takes into consideration both human and technological limitations and possibilities to create the most intuitive conversational user experience possible. In other words, conversation design is about creating conversations between human and machine that feel as natural as possible.    Designing a chatbot is a step-by-step workflow that conversation designers iterate on.   Some questions to consider before designing a chatbot are: Who are your customers? How should they be addressed? What are their needs and pain points while using the bot?  What are the questions your bot should be able to answer and what are the customer's goals behind them? We call these use cases. For instance, a use case can be ordering food or book a train ticket.  What is your bot's persona like? To make the experience as pleasant as possible, your chatbot's personality should be consistent and suited to your customer profile.    Conversation between chatbot and customer       An important part of conversation design is imagining what interactions with customers will look like and designing your bot to handle different scenarios. The happy flow is the ideal interaction that happens when the conversation goes the way the bot designer expects it to. This means the customer gets what they need without the bot running into any errors or dead ends. Happy flow But good conversation design also anticipates and covers potential edge cases. Just like conversations between humans, conversations with a chatbot can jump from one topic to another and customers tend to ask all sorts of things from a bot. So what happens if they start asking questions the bot wasn't designed to answer? How will your bot be able to handle that in the most natural way?   Due to technological limitations, a chatbot will never be able to cover all topics. Therefore, a key point in conversation design is to balance out the bot's use cases (what the bot can talk about) versus the edge cases (topics that are out-of-scope but still need to be handled elegantly).    Happy flow with edge cases Note that conversation design is also an iterative process. Once your bot is live, it is important to monitor and update it depending on how your customers interact with it.    Once you have a clear view of how your bot should behave, Engage allows you to build it intuitively and without any coding skills.       ... View more

A chatbot is an automated conversational agent that uses artificial intelligence to understand and respond to customers' requests in human language. It can take the form of a bot that writes to customers over chat (chatbot) or one that talks to them over a phone call (voicebot), although chatbot is the commonly accepted term for both. A retail chatbot in action Chatbots are a great tool that let your business automate simple tasks like answering frequently asked questions, getting a customer's contact details, or redirecting them to the right support agent. So with the chatbot's help, your team can focus on the important tasks!   Engage allows you to build your bot without any programming knowledge, in a user-friendly and intuitive way. You can use a template that can help you get started faster or build a bot from scratch. Either way, setting up a basic chatbot will take just a few minutes.   🎥You prefer to watch a quick hands-on tutorial? Check out our new bot builder video tutorial!     See also: Chatbot design workflow Conversation design intro Canvas overview ... View more