The role of triggers in workflows
Triggers determine “who” sends input into the Digital Expert workflow and “when” it happens. A single Digital Expert can bind different types of triggers to cover different business entry points. Typical entry points include:- Chat Trigger: User input from the in-platform chat interface.
- Lark Trigger: Input from Lark message events.
- Schedule Trigger: System input from Cron-based scheduled tasks.
Trigger lifecycle
Triggers typically include three phases:- validate: Check whether trigger configuration is complete and valid.
- publish: Register trigger capabilities when publishing or activating a workflow (such as registering callbacks or creating scheduled jobs).
- stop: Release resources when disabling a workflow (such as deleting scheduled jobs or unbinding external integrations).
- replay_publish: Replay publish logic to restore trigger capability automatically.
- skip: Skip auto-recovery and continue through runtime events or external systems.
Chat Trigger Node
When building AI workflows for chatbots and other chat interfaces, use the chat trigger node.Node Parameters
A list of parameters for defining a pre-input form for the digital expert conversation interface, allowing users to enter necessary information before starting the conversation.Related docs
- How a Digital Expert connects to different conversation channels via triggers: Chat Channels
- Trigger plugin implementation details (including Lark and Schedule): Plugin Trigger