In automated systems, item identification is commonly achieved through camera vision, barcode readers, or RFID scanners. However, certain applications, such as identifying layup molds from a diverse range, face challenges with these technologies. This is particularly true for large, manually transported items that share similar profiles, which complicate camera vision systems, while the lack of suitable surfaces eliminates the use of barcode readers. Furthermore, implementing high-cost RFID scanners is impractical due to low flow volume, and retrofitting molds with compatible RFID tags is unlikely. This paper introduces a novel method for identifying layup molds or item IDs at a robotic station, eliminating the need for machine vision, barcode readers, or RFID scanners. Instead, the system operates by making contact at strategically selected points before executing the appropriate routines for cleaning the mold or processing the item. The optimization of touchpoint locations and counts significantly reduces identification time. Unlike the authors’ previous work, which required operator’s input for item ID verification, this approach is fully automated, requiring no human intervention. The methodology employs a modified ‘set covering’ optimization model to mathematically analyze geometric variations among items, thereby minimizing the number of touchpoints the robot visits to accurately determine the ID of the mold or item.