| View supported editions. |
| Einstein Bot Component | Description | Example |
|---|---|---|
| Dialog | Dialogs are conversation snippets that control what your bot can do. You can associate every dialog with a dialog intent, which can be trained to understand variations in customer input. During a conversation with a customer, your bot moves between several different dialogs. Each dialog handles a portion of the conversation. For example, Welcome, Main Menu, Order Status, Location and Hours, and Transfer to Agent are individual dialogs that a customer can experience as part of a single conversation with your bot. |
A dialog for “Chat with an agent”... Customer: “Transfer me to an agent.” [or any variations on this phrase] Bot: “No problem. Hang on: I’ll connect you with an agent.” |
| Entity | An entity represents a type of data that you want to collect from a customer. We provide the following system entities: Boolean, Currency, Date, DateTime, Email Address, Number, Object (standard or custom), Phone Number, and Text. You can create your own custom entities as needed. |
Order number Email address |
| Intent |
Intents are the customer's reasons for interacting with your bot. For example, booking a flight, changing a flight, getting store hours. Associate intents with your dialogs. Then train the bot to create an intent model that your bot can use to understand intents. If your customers interact with your bot by typing a message in the chat window, use intents to help your bot understand what they want. If your customers interact with your bot using only menus or buttons, rather than free-form text input, intents aren’t needed. |
Customer: Where’s my order? |
| Named Entity Recognition (NER) | NER identifies data inside of bot conversations (in other words, an entity) and maintains the format of that data, so that the data is recognizable and usable across Einstein Bots. |
Customer: I need to fly to SFO next week. NER: Extracts “SFO” as an airport code and “next week” as a date. |
| Natural Language Processing (NLP) | NLP is how your bot processes the free text that your customers enter into your bot. NLP combines ideas and methods from computer science, linguistics, and AI to train your bot to identify the literal, grammatical, and contextual meanings of your customers’ inputs. | |
| Natural Language Understanding (NLU) | NLU is a substep of NLP that uses the free text your customers enter into your bot to help your bot decide what action to take. In other words, it uses customer inputs to identify an intent, which tells your bot to perform a useful task for your customer. |
Start an action, such as look up an existing case Transfer to an agent |
| Utterance |
Utterances are common conversational phrases that represent the ways your customers describe an intent. Utterances are the building blocks of intent models, which is why they’re sometimes referred to as intent data. A single intent contains many utterances, because there are many ways to describe the same intent. Utterances are instances of customer inputs. A customer input becomes an utterance when it’s classified into an intent, although sometimes these terms are used interchangeably. You can add utterances to an intent manually, or you can create an Einstein Intent Set to load many utterances at once. |
Intent: Transfer to agent Utterances:
|
| Variable | A variable is a container that stores a specific piece of data collected from the customer or output from Salesforce. Since variables are containers of information, they can be used within dialog actions as both inputs and outputs and can be inserted as part of the text in messages. | If your bot books a flight for a customer, you can create the entity, Airport Code, with two variables, Departure Airport Code and Arrival Airport Code. |