Flow · 02
The operational work, handled.
Flow is the execution layer beneath Pulse. AI agents run automatically each cycle — reading your trading data, applying your strategic anchors, and surfacing structured buying and merchandising decisions for your approval. Approve, defer or override each one, and the decision is kept with its reasoning.
How it fits
Pulse tells you what to decide. Flow handles what follows.
The Pulse brief surfaces the insight — sell-through tracking below plan, aged stock building, the reorder window on your top styles closing this week. That intelligence belongs at the executive layer. The analysis that supports acting on it belongs somewhere else.
The agents do that analysis automatically. Every agent runs off the same trading data that feeds Pulse, applies your stated margin posture and strategic anchors, and produces a structured brief for the relevant buyer or planner. No manual exports. No spreadsheet rebuilds.
The executive sees a summary in Pulse. The buying team gets the full detail from Flow — and approves, defers, or overrides each recommendation. Both happen automatically, every cycle.
Constrained by design
Every recommendation is constrained by your stated strategy.
A margin-protection retailer gets different markdown recommendations than a growth-led retailer — even with identical sell-through data.
A retailer committed to full-price integrity will see that reflected in how clearance recommendations are staged. The strategic anchors you set in onboarding aren't just context for Pulse. They are the decision rules every Flow agent operates within.
The agents don't just run numbers. They run numbers within the constraints of what your business has said it stands for.
The agents
Three agents, running on your cadence.
Built for: Fashion and apparel retailers
For retailers running a test-and-reorder buying model, the business depends on catching winners early and cutting losers fast. The Style Action Agent reads every active style each week — sell-through, cover, age and reorder window — and tells the buyer which styles to reorder and which to mark down before the window closes.
Output
Reorder candidates ranked by urgency with window close dates and recommended quantity. Clear list with margin impact at each markdown depth.