Challenge
Sumatec is a Colombian industrial distributor serving customers across categories where availability matters: tools, hardware, MRO products, safety items, electrical materials, and industrial supplies. The company operates a national inventory footprint with 80 warehouses, 50,000+ SKUs, and more than 3 million inventory records moving through planning logic.
Before Soberan SAI, the demand-planning process depended on a legacy planning vendor and a heavy Excel layer. Planners still had to export, reconcile, and interpret large files to decide which SKUs should be bought, which should move from another warehouse, and which records should be ignored because they had no demand signal.
The implementation had to respect Sumatec’s real planning data: product code, product name, destination warehouse, available inventory, average daily consumption, coverage days, replenishment frequency, lead time, minimum order quantity, packaging multiple, supplier NIT, pending customer orders, stock in transit, product line, and origin type.
Solution
Soberan built SAI as a custom demand-planning workspace on top of the Soberan operating layer. The planner works from one page of glass for replenishment: SKUs, warehouses, demand signals, stock, pending orders, stock in transit, purchase rules, and transfer options are visible in one workflow instead of scattered across vendor exports and spreadsheets.
The data pipeline consolidates data from Jaivana, Sumatec’s ERP, product and catalog context from the Magento PIM, warehouse metadata, demand history, stock positions, and replenishment parameters. It processes millions of records into a planner-ready model so the team can evaluate the right SKU x warehouse rows without rebuilding the dataset manually.
The recommendation engine evaluates each destination row, ignores products without average daily demand, calculates need from daily consumption, lead time, target coverage, safety stock, current stock, pending orders, and stock in transit, then rounds purchase quantities to MOQ and packaging rules.
Before creating a purchase order, SAI searches the other warehouses for safe surplus of the same SKU. If internal stock can cover the need, it recommends a transfer. If not, it recommends a purchase. If both are needed, it produces a split recommendation with source warehouse allocations.
Results
Sumatec’s planners now work from a single demand-planning surface instead of using Excel as the decision cockpit. A selected warehouse view can show thousands of planning records, while the national warehouse model gives context for where supply can move before new purchases are created.
Approved recommendations are submitted to Sumatec’s ERP as grouped purchase orders and warehouse transfers. SAI also writes a traceability record for every transaction, including recommendation-versus-approval audit status, overrides, request payloads, response payloads, document numbers, execution actor, and submission timestamp.
The result is not only a cleaner planner screen. It is a governed operating loop: data arrives from ERP, PIM, warehouse, demand, and inventory sources; Soberan computes what to replenish or transfer; humans review and approve; Sumatec’s ERP receives the execution documents; and the audit trail records what changed.
- warehouses across the country
- 80
- inventory records processed
- 3 million+
- SKUs managed
- 50,000+