A conversation with Bixby is a back-and-forth between Requests made by the user and Moments introduced by Bixby. These Moments make conversations feel comfortable and natural, and help the user get the best possible result to their Request.
The Bixby conversation binds capsule models together in a way that reflects an actual conversation with an end-user. It also defines the pattern you, the developer, should follow when designing your capsule's user experience.
Users can make a request by voice (hands-free mode) or typing (hands-on mode).
Additionally, if the users are in a Moment, they can follow up by tapping on a component on the screen.
In a Bixby conversation, a response is referred to as a Moment. A Moment is composed of three parts, which have distinct sections on a device's screen:
Bixby's dialog is a key aspect of a Moment and sets the context for the rest of the Moment. On a device with a screen, this is displayed in the conversation zone. For more information, see Refining Dialog in the Developers' Guide.
Views render the content of a Moment. The Bixby platform uses Views to scale your content across devices. On a device with a screen, Views are displayed in the content zone.
Depending on the type of Moment, you can include a Conversation Driver or an Action Button to help move the conversation forward. On a device with a screen, Action Buttons and Conversation Drivers are displayed in the action zone.
The following tutorial goes through various Bixby Views components that you can use in your capsule.
When a user makes a request, Bixby is designed to get the user to a result as quickly as possible. However, along the way it might seek input and confirm choices when required. Bixby conversations are made up of a series of these Moments: Results, Inputs, and Confirmations. This design helps the conversation flow smoothly and efficiently, and to stay focused on delivering a great result to the user's question or command:
In a Result Moment, Bixby completes an action or presents information that delivers on the user's goal. This is the most commonly experienced Moment for users. A Result Moment can be in the form of a list of results, a single result, or the receipt of a transaction.
In an Input Moment, Bixby collects and clarifies all the information it needs to complete a user's request. Once all of the necessary information is provided, the conversation moves to the next step.
In a Confirmation Moment, Bixby asks the user to review critical information before taking an irreversible action. This Moment is an opportunity for users to review all of the choices and decisions they have made (or choices that Bixby assumed based on selection or preference learning), before an irreversible action or transaction is completed.
The following are some archetypal conversations from a design standpoint.
For more information on how to handle conversations, including how to handle prompts and follow-ups, see the Guiding Conversations Developers' Guide.
The simplest Bixby interactions only need a Result moment. It's a user's question or command, followed by Bixby delivering the result. Think of straightforward requests like "What's the weather today?" or "Set an alarm for 7am tomorrow morning."
User: "What's the weather like on Venus today?"
Bixby: It's 864 degrees on Venus right now. (Result Moment)
Often users make requests that are open-ended. You could present choices in a Result Moment or quickly clarify the user's intention with an Input Moment. In these conversations, Bixby needs at least one follow up question (an Input Moment) in order to have all the information necessary to deliver a result.
User: "Show me the weather in Coronae City."
Bixby: Did you want the Coronae City on Venus or Saturn? (Input Moment)
Bixby: Today in Coronae City, Venus, it will be 875 degrees with a chance of meteor showers. (Result Moment)
For a more complex example, see the Simple Search Walkthrough Sample Capsule.
Bixby is designed to handle complex, multi-step requests that are about getting things done. This can involve seeking clarification and confirming all the choices the user has made before finally completing the request. These are the most complex Bixby conversations, and require Bixby to both gather information (Input Moments) and confirm a request with the user (Confirmation Moments) before executing an irreversible action like completing a payment, booking a table at a restaurant, or sending a text message.
User: "Book a massage at Venus Space Spa for tomorrow."
Bixby: Venus Space Spa has available appointments at 7am, 8am and 9am. Which one would you like?
Bixby: Okay, I have a massage appointment at Venus Space Spa for tomorrow at 8am. Your total will be $289 galaxy coins. Ready to book it?
(Screen confirms all the information. User taps the "Make a Reservation" action button.)
Bixby: Done. Check your inbox for a confirmation email.
For a more complex example, see the Basic Cart Transactional Sample Capsule Walkthrough.
Users are in one Moment at any time.
Unlike messaging and email, there is no need for the user to go back through a conversation history with Bixby. That's why we use Moments — to ensure the conversation is progressing toward the user's goal.
A Bixby conversation defines how Moments occur.
A conversation with Bixby is defined by how all the Moments can occur within a conversation for a particular capsule. Moments never appear out of order.
Bixby rewards specificity.
The beauty of a conversation is that these moments are strung together depending on what the user asks Bixby to do. The more specific a user is, the further along they go in a conversation.
Every Moment is an opportunity to build a relationship.
Getting to know a user might take time, but each interaction you have with Bixby informs a larger narrative that helps users do things more easily and quickly as Bixby learns.