Glossary
This page defines the terms used across Flametree and these docs, grouped by area and alphabetical within each group. For a guided introduction to the core concepts, read Key concepts first.
Bold words inside the definitions — for example Agents or Conversation Results — are labels you see in the portal.
Workspace and people
Your organization's space in the portal, the people who work in it, and the people your agents talk to.
| Term | Definition |
|---|---|
| 360 View | The screen that gathers everything known about one customer: profile details, conversation history across channels, and collected data. See 360 View. |
| API key | A tenant-level key that authenticates API calls made on behalf of your tenant. A Tenant Admin copies it on the Tenants tab of Settings > Identity. Treat it like a password. |
| Customer | The person on the other side of the conversation. Customers never sign in to the portal — they talk to agents through a channel. |
| Human agent (operator) | A person who takes over conversations the AI cannot or should not finish. Works in the Human Agents workspace, typically with the Tenant Operator role. |
| Operator schedule | The working days and hours of a Tenant Operator. When set, the operator receives handovers only during those hours. Configured in Settings > Identity. |
| Tenant | Your organization's workspace: all agents, knowledge sources, channels, campaigns, users, and billing in one place. Created automatically when you register. |
| User | A person who works inside a tenant. Every user has exactly one role — see Roles and Permissions. |
Agents and configuration
Creating, configuring, and testing AI agents — see AI Agents.
| Term | Definition |
|---|---|
| Advanced mode | The expert configuration view of an agent, where settings are edited directly instead of through the Simple mode steps. See Advanced mode. |
| Agent status | The lifecycle state shown for every agent: Stopped, Starting, Running, Stopping, Error, or Creating. An agent answers customers only while Running. See Agent statuses. |
| AI agent | An LLM-powered assistant that talks to your customers over text or voice channels. Created and managed on the Agents screen — see Agents. |
| Call Analytics | An agent type that does not chat live: it analyzes recorded calls you import and extracts structured insights. See Call Analytics. |
| Inbound Agent | An agent type that replies to messages customers send first — website chat, messengers, email, or inbound calls. |
| Outbound Agent | An agent type that starts the conversation with your contacts, typically as part of a campaign. |
| Playground | The built-in chat for testing an agent before connecting real channels. Available while the agent is Running. See Playground. |
| Simple mode | The step-by-step agent configuration view with the Instructions, Knowledge Sources, Channels, and Conversation Results tabs. See Simple mode. |
| Tool | An extra capability an agent can call during a conversation, such as an external API. Managed under AI Agents > Tools. |
| Wizard | The guided agent-creation flow that builds a first agent configuration from a few questions. See Wizard. |
Conversations and sessions
Live dialogs and their records — see Sessions and Human Agents.
| Term | Definition |
|---|---|
| Billable session | The billing unit: one 24-hour window of conversation with one customer. A single long conversation can count as several billable sessions. See Profile and Billing. |
| Conversation Result | A structured field the agent fills in during a conversation, such as a name, phone number, or selected service. Defined on the agent's Conversation Results tab; collected values appear with the session and in 360 View. |
| Copilot | AI-drafted reply suggestions shown to a human operator. Drafts are never sent automatically — the operator reviews, edits, and sends them. In the chat they carry the Draft from Copilot label. |
| Handoff (handover) | The transfer of a conversation from an AI agent to a human operator — triggered manually, by the agent's handoff conditions, or by a customer request. |
| Private note | An internal note inside a conversation, marked with the Private note label. Visible to your team, never to the customer. |
| Quick replies | Button options an agent offers under a message so the customer can answer with a tap. |
| Session | One conversation between a customer and an agent (or human operator): the full message history, the parameters it started with, the collected results, and a final status. Reviewed on the Sessions screen. |
| Session parameters | The context a session started with — for example, variables passed by the web widget or campaign UTM tags. Shown in the Parameters & results panel. |
| Session results | The Conversation Result values collected in one session. Shown in the Parameters & results panel. |
| Session status | The outcome a session receives when it ends, used in the Session Filters dialog. See Session statuses. |
| System Logs | The session lifecycle log: opening, saving, closing, and tool calls. Opened with the Logs button under a session's chat. |
| Trace Logs | The record of the LLM calls behind a session — full prompts and model responses. Available on paid plans. |
Knowledge
What agents use to answer accurately — see Knowledge Sources.
| Term | Definition |
|---|---|
| FAQ | A named collection of question-and-answer (Q&A) pairs. When a customer asks a similar question, the agent answers with your approved wording. Managed on the FAQs tab of Knowledge Sources. |
| Knowledge Base (legacy) | The previous knowledge system, still visible in older tenants. New setups should use Knowledge Sources — see the migration guide. |
| Knowledge Source | A reusable unit of knowledge an agent draws on: a FAQ, a Plain Text block, a Website, a File, or a YouTube video. Lives in the tenant-wide library at AI Agents > Knowledge Sources. |
| Plain Text | A knowledge source type: a single named block of free-form text. See Knowledge source types for limits. |
| Unanswered questions | Customer questions the agent could not answer from its knowledge, collected on the FAQs tab so you can turn them into Q&A pairs. |
| Usage Count | How many times a knowledge source was used in agent answers. Shown in every source table. |
Channels
How conversations reach your agents — see Channels.
| Term | Definition |
|---|---|
| Channel | The pathway customers use to reach your agent: the web widget, WhatsApp, Telegram, email, voice telephony, and others. Connected in Settings > Channels and attached to agents on their pages. See Supported channels. |
| Web widget | The chat window you embed on your website. Enabled and configured on the agent's page, not in Settings > Channels. See Web Widget. |
| WhatsApp template | A pre-approved message that starts an outbound WhatsApp conversation. The first message of an outbound WhatsApp session carries a WhatsApp template badge. |
Campaigns
Outbound communication at scale — see Flows.
| Term | Definition |
|---|---|
| Campaign | An automated outbound communication process: you upload participants, and a visual flow decides who is contacted, when, and on which channel. Built in Campaigns > Flows. |
| Flow | The visual workflow of a campaign: triggers, conditions, and actions that decide what happens to each participant. |
| Participant | One contact inside a campaign — the person a flow decides to contact (or not). |
Related pages
- Key concepts — the same core terms with more context
- Portal tour — where each concept lives in the portal
- Roles and Permissions — what each user role can see and do
- Agent statuses and Session statuses — the status values behind the terms above