A Pipeline is a sequence of named stages with a defined purpose — sales (Lead, Qualified, Proposal, Won, Lost), onboarding (Signed, Setup, Training, Activated), support (New, In Progress, Resolved). You can run unlimited pipelines in parallel; a contact can sit on multiple boards at once (a Won customer can simultaneously be at "Training" in onboarding and "New" in support). Each pipeline has its own stages, automations, scoring rules and forecast settings.
Stages are configurable as much or as little as you need. Each stage has a name, color, an optional WIP limit (no more than 20 deals in "Proposal" at once), a target duration (deals over 14 days here are flagged stale), and entry/exit automations. Entry automation is "fire a flow when a deal enters this stage" — auto-send the proposal template when a deal moves to "Proposal Sent". Exit automation handles cleanup or notifications. No manual orchestration; the pipeline is a state machine you operate visually.
Drag-and-drop is the primary interface, but every stage transition can also happen programmatically — from a flow node, an API call, an AI Studio tool, or an auto-rule. This dual-mode is critical: humans drag deals when they have context, automation moves deals when signals fire. A deal that has been silent for 21 days in "Proposal" can auto-move to "Stalled" via rule. A new inbound from a previously-Lost contact can auto-move back to "Lead — Re-engaged". The board reflects the truth the conversation is creating.
Lead scoring runs in the background per pipeline. Define rules — +20 for "title contains CTO", +30 for "company size > 500", -10 for "no reply in 7 days" — and every deal carries a live score. The board sorts within each stage by score so reps work the highest-value cards first. Scores feed segments ("hot deals: score ≥ 80") and flows ("if score crosses 90, alert AE in Slack"). The forecast at the bottom of the board rolls up score-weighted deal values per stage to project pipeline coverage.