Skip to main content
This page defines the terms used throughout these docs and inside Swan itself. Each concept has one canonical name — where the product UI or your team uses a different word for the same thing, that’s noted as an alias. Trigger An automation: a set of instructions, the toolkits Swan is allowed to use, and a source that starts it (a schedule, a webhook, a CRM event, an app event, a website visitor signal, and more). See how triggers work. Sequence A multi-step outreach flow sent to a contact — a series of email and/or LinkedIn steps with waits and branching between them, sent through one inbox per channel — an email inbox and/or a LinkedIn inbox, which can belong to different teammates. Every sequence belongs to one Campaign. See how sequences work. Campaign A business initiative that groups sequences, such as “Q4 fintech outbound” or “Conference follow-ups”. Every sequence belongs to exactly one Campaign. Triggers and chat create sequences; Campaigns file them, so one trigger can feed several Campaigns. Each member has a personal default Campaign, and a trigger can have its own default. Older Swan docs and messages sometimes use “campaign” to mean a sequence. See Campaigns. Task An item Swan creates on the Desk for a human to look at or act on — for example, a message waiting on your approval, or something Swan flagged as needing a decision. Internally called a “review item,” but always referred to as a task. See Desk and tasks. Run / execution One pass of the agent doing work — a chat reply, a trigger firing, or a sub-agent working on one entity. The UI may label this a “job.” See executions and history. Inbox A user’s connected email or LinkedIn account, used to send messages and receive replies. Each inbox is one channel, so your email and your LinkedIn are two separate inboxes. A sequence selects one inbox per channel, which can belong to different teammates. See inboxes. User A person in your Swan workspace. Each user can connect multiple email and LinkedIn inboxes; each inbox has its own connection, writing instructions, and sending limits. Credits The unit Swan’s work is metered in, drawn from one org-wide balance. On v1 (per-action credits), most tool calls and trigger fires have a fixed credit price. On v2 (data and intelligence credits, rolling out since October 2026), you pay data credits for data lookups, at fixed public prices, and intelligence credits for Swan’s work, which depend on how complex and how long that work is. See credits. Org knowledge Structured information about your business that Swan reasons from: your value proposition, competitive positioning, ICP definitions, and personas. Distinct from memory (see below). See org knowledge. Memory Freeform notes Swan keeps about your org, contacts, or ongoing work — less structured than org knowledge, used for context that doesn’t fit a fixed field. See memory. Skill A reusable set of instructions that tells Swan how to approach a specific kind of task, so it behaves consistently each time that work comes up. See skills. Library skill A public, ready-made skill Swan can install as a copy into your organization or your personal skills instead of creating the same instructions from scratch. See skills. Desk The screen that surfaces everything Swan wants a human to see or act on: tasks, outreach awaiting approval, and execution activity. See Desk and tasks. Toolkit A group of related tools Swan can use — for example, everything needed to work with HubSpot, or everything needed to manage tags. Toolkits come from three places: built into Swan, connected through the third-party integration layer, or added as a custom MCP server. Your org can enable or disable individual toolkits and set who needs to approve which of their actions. Connected account A third-party login (a CRM, an inbox, a LinkedIn profile, a Slack workspace, and so on) that you’ve authorized Swan to act through. A connected account is what makes a related toolkit or inbox usable. Target market / ICP segment The definition of the kind of company you sell to — the firmographic and behavioral criteria Swan uses to identify a good-fit account. Configured as part of org knowledge. Persona A defined buyer type within a target market — the role, seniority, and characteristics of the person Swan should be trying to reach at a good-fit company. Funnel stage Where a company sits in your pipeline (for example, an early awareness stage through to a closed stage), tracked on the company record and used to reason about what to do next. Tag A label you or Swan can apply to a company to mark it for a purpose (for example, to include or exclude it from something), separate from its funnel stage. List A saved set of companies grouped together, used to scope which companies a trigger, sequence, or query applies to.

​
Common questions

Is a “job” the same thing as a “run”? Yes — some parts of the UI say “job,” but it refers to the same thing as a run or execution. Is a task the same as a review item? Yes. “Review item” is the internal name; the product always shows and calls it a task. What’s the difference between org knowledge and memory? Org knowledge is structured (value proposition, competition, ICP, personas) and drives how Swan reasons about fit and messaging. Memory is freeform notes for context that doesn’t belong in a fixed field. Is a sequence the same as a trigger? No. A trigger is what starts automated work; a sequence is the outreach flow a contact is enrolled in, which a trigger (or a chat conversation) can create. What’s the difference between a tag and a funnel stage? A funnel stage tracks pipeline position (one value at a time); a tag is a free-form label a company can have zero, one, or several of.