Plain-language definitions for the terms you’ll meet across SUPERWISE Chat. Terms are grouped by where you’ll run into them, and listed alphabetically within each group. Where a term has a close relative, the See also note points you to it.
For deeper reading, see The surfaces, Plans and what they include, and the FAQ.
| Term | Definition |
|---|
| Conversation | A single thread of back-and-forth between you (and optionally teammates) and the AI. Conversations are saved automatically, and you can rename, pin, archive, search, or export them. |
| Channel | A persistent, multi-person conversation built for teams. Channels support threaded replies, pinned messages, and a moderator role. Anyone in the channel can talk to the AI, and you can address a specific assistant within it. See also: Direct Message, Thread. |
| Direct Message (DM) | A private conversation with one or more teammates where the AI is still available to help. A DM can be promoted to a Channel when the group grows. See also: Channel. |
| Shared Conversation | A conversation you’ve opened up to specific teammates so they can read along or take part. Everyone sees messages in real time, and each person’s messages are clearly attributed to them. You control who has access. See also: Channel. |
| Folder (Project) | A named workspace that groups related conversations together, along with a shared Assistant setup and Knowledge. New conversations you start while a Folder is open are filed into it automatically. Some surfaces call these Projects — they’re the same thing. |
| Thread | A focused reply branch off a message inside a Channel, so a side discussion doesn’t crowd the main flow. |
| Composer | The message box at the bottom of a conversation where you type, attach files, switch modes, and send. |
| Pin | Marking a conversation (or a message in a Channel) so it stays at the top for quick access. |
| Archive | Moving a conversation out of your main list without deleting it. Archived conversations stay searchable and can be restored at any time. |
| Term | Definition |
|---|
| Assistant | A reusable setup the AI runs against: its instructions (a system prompt), its personality and style (a persona), the Knowledge it can draw on, and the Guardrail Pack that governs it. You pick an Assistant for a conversation, and you can build or edit your own. See also: Persona, Knowledge Collection, Guardrail Pack. |
| Persona | The tone, length, and formatting style an Assistant uses when it replies — for example, concise and formal, or friendly and detailed. |
| Interaction mode (tier) | How much thinking the AI puts into a single answer. You pick a mode in the composer before you send: Interactive (fastest, direct answers), Thinking (the default — reasoning plus search), and Research (deeper search, synthesis, and citations). A fourth mode, Thought Partner — the most thorough, with full search, reasoning, and tools — is available depending on your plan and organization’s settings. See also: Plans and what they include. |
| Thinking view | The expandable panel that shows, step by step, what the AI is doing to build your answer — understanding your intent, checking safety, gathering and filtering information, and composing the response. |
| Citation | A numbered marker in an answer that links to the exact source the AI drew from. Click it to see the source title, a snippet, and a link, so you can verify where an answer came from. See also: Knowledge Collection. |
| Coverage notice | A note that appears when your linked Knowledge had nothing relevant for a question, so the AI answered from general knowledge instead. It tells you when an answer isn’t grounded in your documents. |
| Quick action | A one-click button under an answer — such as Explain more, Refine this, or Try a different approach — that sends a common follow-up for you. |
| Term | Definition |
|---|
| Knowledge Collection | A curated set of documents the AI can search and cite when it answers you. Collections come in scopes: Personal (private to you), Folder (shared with the members of a Folder/Project), and Tenant (available to everyone in your organization). See also: Citation, Folder. |
| Personal Knowledge | A Knowledge Collection that belongs only to you. It stays private — not even an administrator can read it. |
| Folder Knowledge | A Knowledge Collection shared with the members of a Folder/Project. See also: Folder. |
| Tenant Knowledge | A Knowledge Collection available to everyone in your organization. |
| Attachment | A file you add to a single message so the AI can use it for that turn. A useful attachment can be promoted into a Knowledge Collection so it’s reusable later. |
| Web Crawl | Adding the contents of a public website into a Knowledge Collection so the AI can search and cite it. |
| Web Search | Letting the AI look things up on the public web for an answer, with the results cited inline. You turn it on or off per conversation. |
| Note | A saved piece of text you keep alongside your conversations — written by you or saved from an answer — that you can organize and reuse. |
| Snippet | A saved answer (or part of one) captured from a conversation for later reference. Availability depends on your organization’s settings. |
| Template (saved prompt) | A reusable prompt with fill-in-the-blank variables, so you can run a repeatable request without retyping it. Availability depends on your organization’s settings. |
| Memory | Facts and preferences the AI remembers to make future answers more relevant. Personal memory is private to you; Organization memory is shared across your organization and managed by an administrator. You can always review, edit, or clear your personal memory. |
| Term | Definition |
|---|
| Guardrail Pack | The set of safety and policy rules an Assistant follows — the checks that keep answers appropriate, protect sensitive information, and keep the AI within your organization’s standards. Your administrator chooses which pack an Assistant uses. See also: Assistant, Governance. |
| Governance | The protections that make SUPERWISE Chat safe to use at work: every answer is checked against your organization’s rules, sensitive content is screened, and a record is kept of what happened. Governance is about trust, not surveillance. |
| Execution trace | The record of how a specific answer was produced — the steps the AI took and the checks it passed. Available to your administrators and compliance reviewers for oversight. See also: Thinking view, Audit trail. |
| Audit trail | The tamper-resistant log of important actions — such as access changes and policy decisions — kept so your organization can review who did what, and when. |
| Consent | Your control over optional ways your data is used (such as memory or analytics). Essential processing needed to run the service can’t be switched off, but you can manage the optional activities. See also: Your privacy. |
| Term | Definition |
|---|
| Tenant | Your organization’s private, isolated space in SUPERWISE Chat. Your conversations, Knowledge, and settings live inside your tenant and are kept separate from every other organization’s. See also: Tenant isolation. |
| Role | What you’re allowed to do in SUPERWISE Chat. Roles range from read-only access up to organization administration, and your administrator assigns yours. Common roles include Viewer (read-only), Member (the everyday user role), Moderator (manages a Channel), and Administrator (manages people, settings, and governance for the organization). See also: Users and roles. |
| Administrator (Admin) | A person who manages settings, people, Knowledge, and governance for your organization or a Folder/Project. |
| Single Sign-On (SSO) | Signing in to SUPERWISE Chat with your existing company login, so you don’t manage a separate password. Set up by your administrator. See also: Identity and SSO. |
| Plan | Your organization’s subscription level, which sets what’s included — the number of Assistants, monthly usage allowance, and Knowledge limits. See also: Plans and what they include. |
