Lane

A lane is a horizontal sub-section inside a pool. Each lane carries a label, typically a role ("Sales rep", "Credit analyst"), a team ("Compliance", "Finance"), or a system ("CRM", "ERP"). Every activity drawn inside a lane is executed by that role, team, or system. When the process moves from one lane to another, the diagram visually shows a handover.

Lanes are pure grouping. They have no BPMN execution semantics. Tokens do not "enter" or "leave" a lane, they just happen to be in activities that are drawn inside one. That means lanes are a pure readability choice. They help humans scan the process for "who does what" without adding complexity the engine has to handle.

• Help when the same process involves three or four clearly distinct roles and a reader needs to see the handovers at a glance.
• Help when an existing friction in the process is lane-crossing: too many handovers, too many approvals, too many context switches.
• Hurt when you have eight lanes and half of them only contain one activity. The diagram becomes a wall of horizontal lines and the roles lose meaning.
• Hurt when the "role" is actually an individual who could change tomorrow. Prefer team or functional names.

LucidFlow infers lanes from the actors named in the source document. When a transcript says "the sales rep qualifies the lead, then compliance does KYC, then the account manager sets up the account", the generator emits three lanes (one per actor) and places each activity in the lane of its performer. You can toggle between "flow view" (no lanes, compact) and "swimlane view" (full lanes) from the toolbar at any time. The ROI report breaks down costs by lane, which is often the most actionable view for a transformation conversation: "our compliance lane is 40% of the cost of this process".