Botfront.
 Github
Authoring conversations
Getting startedFundamental building blocks: forms and storiesYour first dialogueWhat's next?Conversation builderIntents and entitiesEvaluate understandingConversation contextCreate rich responsesDisambiguate user inputSlot filling and formsCreating custom actionsProcessing incoming utterancesPublish and deployImport (export) from (to) RasaAdvanced topics
Monitoring & Analyzing
Configuring channels
On screen guidance
Proactive conversations
Installation
Developers
Administration
Deployment

Getting started

Botfront is a fully featured conversational platform built on top of Rasa with intuitive interfaces to create context-aware virtual assistants.

Fundamental building blocks: forms and stories

Botfront (and Rasa) provide two main building blocks for dialogue authoring. To decide which one to use, one must answer the question: who is currently driving the conversation?

User drives the conversation: user says something, assistant replies.

Situations where the user drives the conversation are implemented with stories. Common examples are Q&A, and chitchat, but it can be longer conversations as well.

Assistant drives the conversation: assistant asks something, user provides information.

When an assistant is requested to perform an action, it may start by asking relevant questions. For example, collecting contact information through a sequence of questions.

Or asking medical or triage questions to someone asking for a medical consultation.

In Botfront (and Rasa), business logic is implemented with forms

In reality, the user is always in charge.

They take the initiative of the conversation, for example by asking a medical appointment, and temporarily delegate the control to the assistant until the request is completed.

Stories and forms work together to achieve that, or to be more specific, stories orchestrate forms. They activate them, pick-up after completion and handle situations where the user interrupts them.

Your first dialogue

In this section we will illustrate with a very simple example how stories and forms work together.

Creating a simple contact form

Our simple form will collect the user’s first name, last name and email address.

[video]

The information collected is stored in slots. Each node of the graph is a question. When creating a node, you must first decide in which slot the response should be saved.

In this simple example the whole user response is stored in the slot, which is the default setting. Which means that if the user says My name is Nathan the assistant will store My name is Nathan in the slot.

Now that we have created the form, we must create a story to activate it.

Orchestrating the form with a story

To activate the form, the user must first express the intent to get in contact. In the following example, the assistant acknwledges the request and activates the form.

[video]

When the form is activated, the assistant takes control and asks questions until the requested information is collected.

When the form is completed, the story picks up.

As you can see in the video you can access the value of the slots with curly braces (e.g. {first_name}).

What’s next?

Now that we have explained the two fundamental dialogue authoring building blocks, you can read further on both stories and forms.

The Conversation Builder and Conversation Context sections explain in depth how stories and slot work.

The Forms and Business Logic section explain forms in depth, including how to orchestrate them in stories.

🖊️ Edit this page on Github