Key concepts
Before you start to work with your service and read the more detailed documentation, review the key concepts of Watson Assistant Solutions.
Skills
A skill is an atomic, reusable program that represents a capability in a specific domain, for example, providing weather forecasts or controlling your IOT devices in your home, such as your thermostat and lighting.
Watson Assistant Solutions provides a set of built-in skills that are pre-trained and ready for use.
To extend the capability of your AI assistants, use the NodeJS skill boilerplate and SDK to create your own custom skills. Add your own custom skills to the set of built-in skills.
In your custom skills, use natural language understanding (nlu) engines, such as Watson Assistant, regexp, or a custom nlu engine, to understand user utterances.
Create multiple AI assistants by grouping skills into skillsets. You can reuse skills across these skillsets.
When a user asks a question or gives a command, the Watson Assistant Solutions core routing component asks each skill in a skillset to evaluate its ability to respond. Using its routing algorithm, the core routing component sends the user request to the skill that is best placed to respond.
Watson Assistant Solutions routing features enhance the conversation within specific skills and between skills. For example, if you ask a user for clarifying information, in your skill code, use the in conversation flag to take priority over other skills. Use context variables to share information within and between skills and to add personalization. Enhance the responses by providing images or detailed text in card objects.
The core routing component uses the Conversation REST API to send evaluation and converse requests to skills. Review the JSON structure topic to understand the expected format of the converse and evaluate requests and responses.
Add custom command and control skills to your assistants that allow users to control IoT devices from an assistant.
Proactivity
You can design your assistant to initiate a conversation with a user or perform a command. Using the Knowledge and Reasoning alpha component, design your assistant to learn about the world around your users. Develop proactive agents that react to a notification change in the world model of a user. Update the world model based on the notifications to continuously learn about users. Proactively alert users users about changes in their world.