Intro

Conversations serve as dynamic interfaces that enable users to engage with Joule. By adopting a conversational approach, these interfaces empower users to navigate and interact with Joule in a more intuitive and natural manner. Users can skip traditional navigation and describe what they need, whether by typing their queries, speaking aloud, or attaching relevant files. Joule actively suggests context-based prompts, searches across the suite, and delivers comprehensive answers with live data and actionable UI elements. This approach not only enhances user engagement but also simplifies the process of obtaining the desired information or performing specific tasks, creating a seamless and efficient communication experience.

Anatomy

Conversations anatomy

1. Left pane conversation toolbar: supports search, filtering, and sorting of conversations, as well as starting a new conversation
2. Conversation actions: provide contextual actions on each conversation item, including managing relations, pinning, renaming, and deleting
3. Conversation header: provides conversation-level controls, including expanding or collapsing the layout, viewing and editing the conversation title, and accessing related objects such as spaces, documents, or contacts.
4. Message thread: displays the sequence of user and Joule messages, forming the core interaction flow of the conversation.
5. Response actions: appear directly below each response and provide contextual controls, including copying content, rating the response, retrying, saving to memory, and accessing insights and sources.
6. Joule input field: enables users to compose and send messages, supporting multi-line input, file attachments, mentions of assistant, and message submission through the send action.
7. Response details (right pane): slides in from the right side of the screen when the user selects "Sources" in the action toolbar. It provides full transparency about how Joule generated its response, covering reasoning steps, referenced sources, and saved memory.

Variations

Full-Screen Conversation

The fullscreen conversation surface gives users a focused, distraction-free space to interact with Joule. Users can interact with Joule not only within the conversation interface but also throughout the entire engagement layer ecosystem, making it available beyond the full-screen view. By highlighting or selecting content in context, users can effortlessly trigger Joule input, which offers quick prompts with contextual suggestions.

Full-screen conversation landing page

Conversations within Jobs

Within jobs, conversations are seamlessly integrated, enabling users to remain engaged with ongoing processes while interacting with Joule. Users can inquire about job progress, interpret system outputs, and intervene when needed, without moving away from the task. This setup allows them to review execution details, clarify results, and make real-time corrections, ensuring they stay in charge of automated workflows while maintaining continuity.

Conversation within a job

Conversations within Spaces

In spaces, conversations help users by collaborating and managing structured tasks. Joule assists by summarizing information, clarifying decisions, and suggesting next steps alongside workspace content. This allows users to work through complex scenarios and make updates without changing contexts, fostering a seamless flow between thinking, discussing, and executing tasks.

Conversation within a space

Behavior and Interaction

Managing Conversations

The conversations area is a dedicated conversation hub where users can start a new chat, browse their full history, and use search and filters to quickly find and reopen any previous conversation so they can resume work with the right context instead of starting from scratch.

Conversations area

Searching and Sorting Conversations

The search and sorting capabilities simplify navigation across conversations and within lengthy threads. Users can find specific topics, business objects, or past decisions without having to manually scan entire histories. This functionality promotes efficient retrieval, making conversations easier to review and reuse.

Searching conversations

Refining Requests with Type-Ahead Suggestions

The type-ahead functionality offers real-time suggestions as users type, taking into account the business context and available data. This helps users reference the correct objects, refine their queries, and minimize input errors. By guiding users toward more precise input, the system enhances both the speed and accuracy of responses.

Input field with type-ahead suggestions

Using the Prompt Library

The prompt library provides a structured collection of reusable prompts and workflows that showcase Joule's capabilities. It encourages users to go beyond straightforward questions by presenting reliable and validated use cases. This feature aids in discovery and encourages deeper engagement with the system.

Prompt library

Using the @Mention Feature

The @mention feature enables users to include specialized assistants in the conversation. By referencing a domain-specific assistant, such as HR or Finance, users gain access to focused expertise while remaining within the current thread. This facilitates more efficient task management and supports complex workflows that span multiple domains.

@mention feature referencing HR assistant