Concepts & Glossary
This glossary outlines the key terminology used across SupportOS. It is written for support managers, team leads, and operators building and managing AI agents.
Workspace & Team Management
| Term | Meaning |
|---|---|
| Organization / Workspace | Your company's account on SupportOS. Everything you create — agents, knowledge bases, actions, phone numbers, and tickets — lives inside your organization. |
| Team Member | A colleague who has access to your SupportOS workspace to build agents, view customer tickets, manage knowledge, or review analytics. |
| Role & Permissions | The level of access assigned to a team member (such as Administrator or Read-Only), determining what settings they can view or edit. |
| Active Workspace | For team members with access to multiple companies, the specific workspace currently selected in the dashboard. |
AI Agents & Lifecycle
| Term | Meaning |
|---|---|
| AI Agent | A virtual support representative customized with your brand instructions, behavior rules, knowledge documents, and actions to assist your customers. |
| Agent Type | The communication format of your agent: Chat (for text-based messaging) or Voice (for phone calls). |
| Draft | Your private, editable working copy in the agent builder. Changes made in the draft do not affect live customer conversations until published and deployed. |
| Version | A saved, permanent snapshot of an agent's configuration at a specific moment in time, with release notes and author details. |
| Publish | Saving your current draft into a numbered version in your agent's history. Publishing stores your changes safely without making them live immediately. |
| Deploy | Activating a specific published version so it begins handling real customer conversations on your live channels. |
| Live / Active Version | The specific version of your agent that is currently interacting with live customers. |
| Restore Version | Loading an older version back into your draft editor so you can review it, edit it, or re-deploy it. |
| Agent Status | Indicates whether an agent is actively serving traffic (Deployed) or offline (Inactive). |
| Archive | Soft-deactivating an agent to remove it from active channels while safely preserving all its settings, versions, and conversation history. |
Deep dive: Agents — Overview and Versioning & Deployment.
Behavior, Scenarios & Guidelines
| Term | Meaning |
|---|---|
| Scenario | A dedicated playbook or workflow that instructs the agent on how to handle a specific customer intent or topic (e.g., "Return Policy" or "Order Inquiry"). |
| Scenario Matching | The automatic process where the AI understands what the customer needs and selects the right playbook to guide the conversation. |
| Rules | Non-negotiable instructions and policies that the agent must adhere to across all conversations (such as privacy guidelines or escalation conditions). |
| Brand Instructions | The core prompt that establishes your agent's identity, company background, conversational style, and overall tone. |
| Personalization | Custom identity settings for your agent, including display name, primary language, operating timezone, and voice profile. |
Scenario vs. Knowledge Base
- Scenario: How to guide a customer through a specific process or workflow (step-by-step actions, tone, and relevant tools).
- Knowledge Base: What information the agent can look up (factual documentation, policies, FAQs, and product details).
See the full decision guide in Scenarios vs. Knowledge Base.
Knowledge Base (Content & Documents)
| Term | Meaning |
|---|---|
| Knowledge Base (KB) | A organized library of documents, FAQs, and articles that your agents can search through to answer customer questions accurately. |
| Knowledge Entry | An individual document, uploaded file, or written article stored within a Knowledge Base. |
| Document Processing / Indexing | The automated preparation step where SupportOS reads and optimizes uploaded content so agents can find answers quickly. |
| Knowledge Attachment | Linking one or more Knowledge Bases to an agent (or a specific scenario) so the agent can reference those documents during customer conversations. |
Deep dive: Knowledge Bases Overview.
API Tools
| Term | Meaning |
|---|---|
| API Tool | An API connection that allows your agent to look up dynamic information (such as checking an order status, account balance, or appointment time) during a conversation. |
| Tool Testing | A preview sandbox in the dashboard that lets you test API endpoints with sample inputs to verify accuracy before enabling the tool for an agent. |
| Tool Attachment | Assigning an API tool to an agent or scenario so the agent knows when and how to perform the lookup for customers. |
Deep dive: API Tools Overview.
Phone Numbers & Voice Telephony
| Term | Meaning |
|---|---|
| Phone Number Provisioning | Acquiring a dedicated phone number directly from the dashboard for your voice agents. |
| Number Assignment | Connecting an active phone number to a voice agent so incoming customer calls are answered automatically. |
| Number Release | Removing an unneeded phone number from your account to cancel its recurring service. |
| Supported Countries & Regions | The catalog of available international countries and regional area codes supported for voice calling. |
Deep dive: Phone Numbers Overview.
Widgets & Integrations
| Term | Meaning |
|---|---|
| Web Widget | An embeddable chat window that you can add to your website so visitors can interact directly with your agent. |
| Helpdesk & CRM Integration | Connecting SupportOS with platforms like Freshchat to unify messaging channels (WhatsApp, Instagram, web chat) and tickets. |
| Human Handoff / Escalation | Seamlessly transferring a customer conversation from the AI agent to a live human agent when specialized help is required. |
Deep dive: Widgets and Integrations.
Conversations, Tickets & Quality
| Term | Meaning |
|---|---|
| Conversation / Session | A single interaction between a customer and an AI agent across chat or phone. |
| Test Preview / Simulation | A sandbox in the agent builder where you can chat with or call your agent to test behavior before deploying. |
| Support Ticket | A structured record generated after a conversation, containing conversation summaries, key timeline events, customer sentiment, and tags. |
| Tag | A categorization label attached to tickets automatically by the AI or manually by team members to classify issues. |
| Call Recording | Audio playback of a customer voice conversation available directly inside the ticket record for quality assurance and training. |
| Agent Evaluation | Automated quality testing suites that run simulated conversations to score agent accuracy, helpfulness, and policy compliance. |
| Manual Ticket | A ticket created directly by a team member for back-office tracking or offline customer requests. |
| Ticket Reference ID | A unique, user-friendly tracking code (e.g., TICK-62YHW) used by team members and customers to reference an inquiry. |
Deep dive: Tickets & Recordings, Tags, and Evaluation Overview.
Billing & Usage
| Term | Meaning |
|---|---|
| Usage | The volume of conversation minutes, messages, and phone numbers consumed by your account during a billing cycle. |
| Invoice | A billing statement outlining your plan, itemized usage charges, and payment status for each period. |
| Payment & Receipts | Records confirming completed payments for account invoices and subscription renewals. |
Deep dive: Billing Overview.
Testing Modes
| Mode | What It Tests |
|---|---|
| Draft Test | Tests the agent using your current, unsaved working edits directly in the builder. |
| Version Test | Tests the agent against a specific published snapshot to verify previous behavior. |
| Live Deployment Test | Tests the exact version that customers are experiencing right now on your live channels. |
Next Steps
Explore the Mental Model to see how Workspaces, Agents, Channels, Conversations, and Tickets connect across SupportOS.