Quick Links
Video Overview
https://vimeo.com/730016175For full screen please view on Vimeo at this link
About Conversational AI
Conversational Artificial Intelligence (AI) refers to technologies, (like chatbots or virtual agents or in the case of Motive, characters in VR), which users can talk to.
They use large volumes of data, machine learning, and natural language processing to help imitate human interactions, recognizing speech and text inputs and translating their meanings across various languages.
Source: https://www.ibm.com/cloud/learn/conversational-ai
Key Components
When you open the AI area of SF, you will see three headings: Agents, Intents, and Keyphrases. These are the components in Storyflow authoring that allow you to create AI interactions:
Agents
The Agent is the part of the system that is processing the language and determining keyphrase and intent matches
Because of this, an Agent is tied to a specific language:
You can see in the example above that there is an English agent and a French agent.
Typically however, a project only needs one Agent (unless there are multiple languages being used). Each Intent Catalog and Keyphrase needs to be connected to an Agent.
Learn more about Agents and how to create them here.
Intents
An Intent is the intention behind the message the AI service receives from the person speaking
An intent-based agent, then, works by detecting this user intent. So, instead of relying on specific input or set of keywords, the agent can identify the meaning the message is trying to convey. (And then offer a relevant, tailored response.)
In the above example, I have an Intent for “Greeting”.
I’ve authored several different ways to convey the meaning of the Intent. In the environment, the learner says, ‘Hello, thanks for coming in today.’ The Intent behind the message is to greet the person/character they are speaking to. The more examples you include in an Intent, the more likely the AI will be successful in identifying the Intent and returning the expected response. So you could add phrases such as “Thanks for coming in today” or “Thanks for joining me”.
Learn more about Intents and how to create them here.
Keyphrases
A Keyphrase is a specific word or phrase that the user needs to say to be recognized
Examples:
Learner says “Help Me” or “Inventory” to trigger this assistance in environment
There is only one right way to say something or you want to enforce the use of one specific turn of phrase
You want to present the user with specific choices in a conversational scenario and have them verbally dictate what they see on a screen to vs speaking freely
Learn more about Keyphrases and how to create them here.