A system using voice recognition is provided to assist in restocking shelving while simultaneously creating an accurate product-to-location mapping of the overall environment. The creation of such a product-to-location mapping enhances many in-store tasks, such as stocking, picking, and cycle counting. The system allows store based task-management software systems to efficiently direct users to precise locations, to reduce product search time, to improve worker travel paths throughout the store to save time, and to free workers to perform other tasks. Workers use a voice-recognition restocking application to record, via voice-entry, the location where each restocked product is stored. The voice capturing of the product-to-location mapping is conducted over time to provide a seamless and sustainable way to maintain product-to-location store maps.