A stock count is a simple test with an uncomfortable result. The system says there are 40 cases of stone fruit on the floor. The count finds thirty-one. The nine cases were not stolen. They were moved, sold, or damaged at some point nobody recorded, and the record has been wrong ever since.
For a producer holding perishable stock, that gap is not an accounting nuisance. It sets the reorder quantity and the markdown decision for everything downstream, and both are now built on a number that is not true. Altinteg, a Portugal-based traceability company, has built its business on closing the gap between what a system believes and what is physically present.
The Count That Does Not Match the System
The company’s own product illustrations show what item-level identification looks like in practice. A single burrata carries the identifier PUG04-D2M. A loaf of artisan bread carries EMR36-B9K. A bottle of red wine carries TOS20-R7F. Each is a demonstration rather than a customer’s stock code, but the principle behind them is the operative one. The unit of truth is the item, not the case or the pallet.
Category-level tracking works acceptably for goods that sit still and keep. It fails with perishables because what an operator most needs to know is which specific units are closest to the end of their shelf life. A case number cannot answer that. Forty cases of fruit received on the same day do not deteriorate at the same rate, and averaging them produces a write-off that arrives as a surprise.
What Changed at the Fruit Producer
Altinteg reports four results from a deployment with a fruit producer. Inventory counts finished 65 percent faster. Shipment preparation improved by 35 percent. Spoilage fell by 20 percent. Profitability rose by 10 percent. The figures are the company’s own, and the producer is not named publicly, so they should be read as reported outcomes rather than independently audited ones.
The order of those numbers matters more than their size. Faster counting is a labor-saving. Better shipment preparation is a throughput gain. Reduced spoilage is recovered product that would otherwise have been destroyed. The profitability figure is what happens when the first three are no longer separate problems. None of it required the producer to sell differently or to buy new stock. It came from knowing, at the item level, what was already in the building.
“We build practical, end-to-end systems that connect physical products with usable data in a way that works in real operations,” said Aliya Pogorelskaya, founder and CEO of Altinteg Technology Solutions.
That phrasing carries a claim about failure as much as success. Most documented disappointments in RFID adoption are not technology failures. A pilot performs in a clean warehouse with trained staff, then stalls on a working line in a perishables distribution center at four in the morning.
The Order the Layers Go In
Altinteg describes its architecture in six layers, and the sequence is the argument. Standards and governance come first, aligned to GS1, so that an identifier means the same thing to a supplier and to a retailer. Digital identity comes second, through serialization and item-level digital twins. Only then do data carriers appear, where labels, optical codes, and RFID tags reside. Detection infrastructure follows, then middleware that filters raw reads into an existing ERP, warehouse, or retail management system. The intelligence layer sits last, handling forecasting and expiration.
Most buyers start at the third layer because it contains the price list. Tags and readers are procurable. Serialization schemes and governance are not, so they get deferred, and that deferral produces a pilot that reads tags accurately yet still cannot tell the finance team anything useful.
The company’s engagement model reflects that. Feasibility and return modeling is the second of its five steps, before any deployment blueprint exists. Altinteg develops alongside European RFID Labs and Universities and aligns its technology with GS1 and the RAIN Alliance, a deliberate choice to build on identification schemes that the rest of the supply chain already recognizes.
The nine missing cases are the whole business case. A producer who cannot reconcile a count is not short of software. That producer is short on a reliable identity for each item, and until the identity exists, every downstream system is confidently processing a number that the shelf disagrees with.