Glossary
Key terms used throughout AgentDesk.
| Term | Definition |
|---|---|
| Operator | The portal role for team members who work on tickets and collaborate with customers (formerly called "Agent"). Not to be confused with AI agents that automate workflow steps. |
| API key | A personal access token that allows programmatic access to the AgentDesk API. Keys have granular scopes and an optional portal restriction. |
| Approval gate | A checkpoint in an AI workflow where the system pauses and waits for a human to review and approve a proposed action. |
| Asset | A URL or endpoint registered for health monitoring. AgentDesk periodically checks assets and can create tickets automatically when downtime is detected. |
| Attachment | A file uploaded to a ticket. Attachments are stored securely and can be downloaded by anyone with access to the ticket. |
| Backlog | The set of tickets assigned to a sprint. Each ticket in the backlog can have an estimate in story points. |
| BigCommerce | An e-commerce platform integration that syncs product and order data into your portal. |
| Burndown | A chart or metric showing remaining work in a sprint over time, helping teams track progress toward completing all backlog items. |
| CSAT | Customer Satisfaction -- a feedback mechanism where customers rate their experience after a ticket is resolved, using stars (0--5) or thumbs (up/down). |
| Custom domain | A branded URL (like support.yourcompany.com) that you configure for your portal instead of using the default AgentDesk subdomain. |
| Customer | A portal role for external users who create and track their own support tickets. |
| Escalation | The process of notifying additional team members when an approval request has not been acted on in time. |
| Image credits | A balance of credits consumed when the AI generates images. Credits are tracked per organisation and replenished each billing period. |
| Incident | A period of detected downtime for a monitored asset, from the initial failure until the asset recovers. Incidents can trigger automatic ticket creation. |
| MCP | Model Context Protocol -- a standardised way for external AI tools to connect securely to your portal. |
| OIDC | OpenID Connect -- the authentication protocol used for Single Sign-On (SSO) connections. |
| Organisation | The top-level account in AgentDesk. An organisation contains one or more portals, users, and billing settings. |
| Org Admin | An organisation role that can manage users and portal settings. |
| Org Owner | An organisation role with full access to all settings, billing, and user management. |
| Portal | A branded workspace for a specific client or team. Each portal has its own users, tickets, branding, and settings. |
| Portal Admin | A portal role that can configure portal settings, manage users, and review AI workflows. |
| Prime Directive | Behaviour guidelines set for the AI agent in a specific portal, controlling what the AI focuses on and how it behaves. |
| Pull request (PR) | A proposed code change on GitHub. AI workflows create PRs for your team to review before any code is merged. |
| Request type | A category for tickets (such as Bug Report or Feature Request) that determines the workflow and Jira issue type. |
| Shopify | An e-commerce platform integration that syncs products, orders, and customer data into your portal. |
| SLA | Service Level Agreement -- a set of response and resolution time targets per priority level, tracked against business hours for each portal. |
| Sprint | A time-boxed period (with start and end dates) used to organise and track a set of tickets. Sprints have statuses: planned, active, completed, or cancelled. |
| Sprint snapshot | A record captured when a sprint is completed, showing committed, completed, and carryover ticket counts and story points. |
| SSO | Single Sign-On -- allows users to sign in through their company's identity provider instead of using a separate password. |
| Tenant | Another term for an organisation in a multi-tenant system. Each tenant's data is isolated from others. |
| Ticket | A support request submitted through the portal. Tickets can be tracked, commented on, and resolved by your team and AI workflows. |
| Transition | A defined status change for a ticket (for example, from "Open" to "In Progress"). Transitions can be restricted by role. |
| Watcher | A user who has subscribed to receive notifications about a specific ticket. Watchers are notified of status changes, comments, and other updates. |
| Webhook | An automated notification sent between systems when an event occurs (for example, Jira sending an update to AgentDesk when a ticket changes). |
| Workflow | The series of statuses and transitions that a ticket moves through. Also refers to AI workflows that automate ticket processing. |
| Workflow mode | The processing strategy chosen by the AI for a ticket: full processing, analysis only, quick fix, or human only. |