See demand ahead of time, set up stock accordingly
Sales history, seasonal effects, campaign calendar and supplier lead times come together in one engine. Kahin, HaROS's demand forecasting engine, analyzes sales, historical data, live stock and traffic. The replenishment, purchasing and production plan comes out automatically.
Six steps from data to plan, a better forecast with every loop
A forecast is not a one-off report. It refreshes every day and corrects itself with actual data.
Data is collected
Sales history, stock, campaign calendar and supplier lead times from all channels are collected in one pool.
one poolDemand is forecast
Kahin forecasts demand by category, location and season. The forecast goes down to SKU × location level.
SKU × locationAnomalies are caught
Unexpected deviation, campaign spikes or demand for out-of-stock items are flagged. They are kept apart from normal fluctuation.
anomaly detectionReplenishment plan is produced
Which product goes to which warehouse, when and how much is determined. Replenishment tasks are generated automatically, before the shelf empties.
automatic replenishmentPurchasing and production are triggered
The system suggests a purchase order for the supplier and a production plan for the floor. The system does the calculation, you give the approval.
purchasing · productionThe model corrects itself
Actual sales and stock return to the model. The forecast becomes more accurate with every loop.
learning loopHistory, forecast and confidence interval on one screen
For every SKU and location you see actual sales, the forecast and the confidence interval together. When the reorder threshold is crossed, the replenishment task is created on its own.
- Actual vs. forecast — deviation visible instantly
- The confidence interval shows the uncertainty, so you decide accordingly
- Automatic replenishment when the reorder threshold is crossed
Illustrative demand forecast view
The forecast is fed with real operations data
The more data the forecast is fed with, the more accurate it gets. HaROS works with the live operations data of 50+ brands.
Annual sales volume processed on the platform.
Product lines processed daily.
The plan is refreshed every day and corrected with actual data.
Forecast along SKU, category, location and season axes.
The inputs, breakdown and outputs of planning
HaROS sets planning up around your operation: at which breakdown the forecast is made, with which inputs, and for which outputs.
Forecast, replenishment and production plan
The system produces the forecast, replenishment triggers itself, and the plan is aligned with finance and production.
The system produces the forecast
Kahin produces an automatic forecast by category, location and season. Anomaly detection flags the unexpected and separates the price and promotion effect.
Replenishment task is created before the shelf empties
When the reorder threshold is crossed, HaROS WMS opens the replenishment task automatically. The picking location does not run empty and excess stock does not build up.
The plan connects to purchasing and production
Purchase order and production plan suggestions come out of the demand plan; suggestions are passed to ERP and approval stays with you.
Planning feeds the warehouse and order flow ahead of time
Planning is not just a report; it lands in WMS as a replenishment task and in OMS as a stock distribution suggestion. Actual sales and stock flow back: one stock, one panel.
Supply Chain Planning
Forecasts demand, catches anomalies; produces the replenishment, purchasing and production plan.
● YOU ARE HEREWarehouse Management
The replenishment task is carried out on the floor: barcode-verified picking, shelf placement and automatic replenishment.
Open page →/span>Order Management
Orders from every channel in one pool — the rule engine fulfills from the best-fit location.
Open page →/span>Frequently asked questions
How much historical data does forecasting need?
Ideally a few seasons of sales data, but you can start with less. The model corrects itself with actual data every day. You do not need to wait for enough data to build up.
Does it change our existing ERP and purchasing process?
No. Planning produces purchasing and production suggestions; your approval process and ERP stay as they are. It works two-way with ERPs such as SAP, Oracle, Logo and Nebim: the suggestion goes to the ERP, the approved order comes back.
How does it separate campaign and season effects from normal demand?
The campaign calendar and promotion plan are given to the model as inputs. Anomaly detection separates campaign-driven spikes from unexpected deviation. This way a one-off campaign is not read as lasting demand.
How does it forecast a new product with no history?
The model derives the first forecast from similar product and category profiles. As the product sells, the forecast is updated with its own data.
Is replenishment triggered automatically?
The suggestion is always generated automatically. You choose whether triggering is fully automatic or with approval. In WMS, the replenishment task is opened before the picking location empties.
See demand ahead of time and set up stock accordingly
Both stockouts and excess stock come from forecasting errors. HaROS ties planning to your operation; your ERP keeps running.
Request a demoWarehouse: HaROS WMS → · Orders: HaROS OMS → · Integrations →