Chat interface
Interface dedicated to conduct conversations with the Elastic AI Assistant.
The chat interface is designed to mimic the natural flow of a conversation, providing a familiar and intuitive medium for users to engage with the AI Assistant, seek assistance, ask questions, and receive guidance.
When to use
Complex queries
When users have specific, complex questions that require a conversational context to understand and resolve
Guided assistance
For step-by-step guidance through intricate workflows or troubleshooting
Personalized interactions
When users seek tailored advice or insights based on their data.
Chat skeleton
The chat interface is structured into the following containers:
- Chat header - it’s purpose is to provide top level controls, which are impacting things outside of the conversation.
- Conversation header - this area displays the name of the conversation and gives quick access to controls which may impact the conversation such as, edit title, change Connector, enabled Knowledge base etc.
- Conversation body - this area is structured as a conversation timeline, where a user can see and interact in a top to bottom, old to new display of messages and events.
- Chat footer - the footer is constructed of three layers which can be used for adding/changing/updating the text input to enrich or change the request which will be made to the assistant: input mode, from text to document, JSON schema mode adding context, and adding prompts.
Chat message types
Loading state
Indicates the assistant is processing the request.
Event state
Markdown formatted message
Query message
Visualisation message
Collapsed messages
Chat message skeleton
The chat message is a dynamic component engineered to support various message states. This offers customisation options that enhance the user experience thoughtfully.
Chat message header - designated section to identify the message's author, with the current setup featuring two participants: the user, and the AI Assistant. On the right side, there's an interactive actions area, providing space to incorporate additional actions.
Chat message body - central area which hosts the message's content, ranging from simple text to complex elements like tables, visualisations or codeblocks.
Chat message footer - positioned at the bottom, this segment contains actions, with options such as feedback on the message's helpfulness, triggers to regenerate the message, etc.