AGENT 01 / 11 · RETAIL & E-COMMERCE · SUPPLY CHAIN & INVENTORY

Demand Forecasting Agent

Projects SKU-level demand forward by blending trailing sales velocity with trend and seasonality — the weekly forecast a demand planner would rebuild from a spreadsheet every Monday, generated instantly and re-runnable against any assumption.

READS FROMPOS transaction log, promotional calendar

MCP SERVERRetail Data Server

MODELRouted to reasoning-tier open model

HANDS OFF TOInventory Optimization Agent, Overstock & Markdown Planner

RUN THE AGENT

Set your assumptions, generate the forecast

This is a live working demo of what the agent produces — adjust the assumptions and it recalculates the weekly forecast, chart and KPIs instantly.

ASSUMPTIONS

Baseline weekly demand 800 units

Weekly trend growth 1.2%

Seasonality amplitude 20%

Seasonal cycle length 12 wks

Forecast horizon 12 wks

—

Peak weekly demand

—

Average weekly demand

—

Total forecast units

—

Forecast swing, peak v trough

HOW IT WORKS

What the agent actually does

1. Pull sales history	Reads 12–24 months of POS transactions and the promotional calendar via the Retail Data MCP server.
2. Detect the pattern	Decomposes trailing sales into a trend component and a repeating seasonal curve.
3. Project forward	Extends the trend and re-applies the seasonal curve across the chosen horizon.
4. Hand off	Passes the SKU-level forecast to the Inventory Optimization Agent for reorder planning and the Overstock & Markdown Planner for excess-risk screening.