When the pickers in Foster & Gallagher’s Peoria, IL, distribution center appear to be hearing voices, it’s normal – not paranormal. Many of F&G’s pickers receive their instructions over special headsets – part of a voice recognition system designed to speed the picking and packing process.
Foster & Gallagher, a $480 million multititle cataloger, has used speech recognition technology in its children’s catalog division, which includes the HearthSong and Magic Cabin Dolls titles, since 1997.
Designed for high-speed picking and packing during the peak holiday season, the voice recognition system lets F&G pickers remain focused on completing tasks while they are receiving instructions and confirmations through their headsets. Workers enter completed actions by voice, which means they have both hands free for picking and sorting merchandise.
For example, a worker preparing to pick products from a specific zone would recite into the headset the SKU and a brief description of the first product to be picked, such as “item number 1145, blue.” The picker would then get confirmation through the headset from a preprogrammed inventory profile system and move on to picking the next item.
According to Nick Licht, manager of information systems and inventory control for Foster & Gallagher, “With the headset, you can do exactly what a scanner would do, except the headset never takes away from your visual focus.” It also cuts down on error rates, he says. Prior to implementing the system, F&G’s children’s catalogs’ peak season throughput was 16,000 orders per day. In 1998, the division peaked at 30,000 orders per day, with the same labor level. And order accuracy has improved 20%, Licht says, while inventory accuracy improved 25%.
Although F&G won’t reveal the cost of its voice recognition technology, such a system with 10 users would likely cost $130,000-$200,000, including hardware and installation, according to Princeton, NJ-based manufacturer Voxware.
Voice or speech recognition has come a long way in recent years, says Dennis Bathory Kitz, a Burlington, VT-based technology consultant. Even five years ago, the speech was sometimes garbled, and the systems recognized only a narrow vocabulary, he says. But within the past 18 months, the technology has improved to recognize more words and to respond more quickly to queries, he claims.