Speech Technology Magazine

 

A Matter of Convenience

A c-store distributor realizes double-digit gains with a voice-driven warehouse application.
By Leonard Klie - Posted Jan 1, 2012
Page1 of 1
Bookmark and Share

Eby-Brown distributes name-brand food, candy, tobacco, health and beauty products, and other items to 13,500 convenience stores in 20 states throughout the eastern and midwestern United States. It also owns and operates a commissary where it manufactures its own line of fresh sandwiches under the Wakefield Sandwich Company label. With more than $4.5 billion in annual sales, the Naperville, Ill.-based company is the third largest convenience store distributor and wholesaler in the country.

Eby-Brown operates more than 1.4 million square feet of warehouse space spread across seven facilities in Wisconsin, Ohio, Illinois, Indiana, Michigan, Maryland, and Georgia. Each one of those facilities has realized significant increases in productivity and order accuracy using the Jennifer VoicePlus warehouse application from Lucas Systems, a Wexford, Pa.-based provider of voice-directed warehouse applications.

The Jennifer VoicePlus system uses proprietary text-to-speech and speech recognition technologies to facilitate conversations between the host computer or warehouse management system (WMS) and the warehouse workers on the floor. Workers interact with the system through headsets connected to voice-enabled mobile computers, freeing their hands and eyes to focus on the job at hand. The system, available in many languages, directs workers to specific locations within the warehouse and tells them how many items to pick from the shelves for each order. The system can verify and respond to the information that workers speak back. Workers can ask the system to speak faster or slower, softer or louder, and can even ask for help and more information so they can do their jobs without stopping to ask a supervisor.

The system not only accepts voice input; data can also be entered into the system through integrated bar code and radio-frequency identification (RFID) scanners.

The Jennifer server-based solution also includes the Engage Web-accessible management services console with configurable user, assignment, route, and task management capabilities, as well as real-time productivity and management reports that can be tailored to exact requirements. It offers managers real-time insight into operations on the warehouse floor, including process status, associate productivity, shorts, and exceptions.

Since Jennifer was first installed, those operations at Eby-Brown are now reporting a 46 percent drop in the number of errors and "sizable gains" in productivity, according to Chris Timmons, executive vice president of operations. Productivity, measured in the number of items picked per person in an hour, is up 33.4 percent for full cases and 40.1 percent for individual items.

Equally important, the Jennifer system gives Eby-Brown additional flexibility to optimize picking assignments outside of the WMS. "This has recently allowed us to increase our shipping volumes without making any additional capital investments in our distribution centers," Timmons says.

The Jennifer system can be used in any number of warehouse activities, including picking and order selection, receiving, put-away and replenishment, loading, returned goods processing, equipment inspection, and quality control. Eby-Brown is currently only using it for picking and packing.

"There are other applications that could benefit from voice," Timmons admits, "but picking was a slam dunk for us. There was a very quick ROI."

Pick-and-pack, he adds, "is the single largest labor expense, so it was an opportunity to reduce costs."

About 600 workers at Eby-Brown use the Jennifer system daily. Warehouse operations at Eby-Brown run 24 hours a day, but only the afternoon/evening shift and overnight shift use voice. The first shift of the day handles receiving and put-away; neither is automated yet.

Prior to the system's installation, Eby-Brown's warehouse was run with paper-based processes and manual work-arounds to address claims and unique customer demands. To say the old way of doing things was inefficient would be an understatement. There were just too many things that couldn't be done properly, Timmons would argue.

A Slow Start

Eby-Brown initially deployed Jennifer on one shift in the candy section of its warehouse in Ypsilanti, Mich., in October 2009, then expanded it to other shifts and departments before rolling it out to the other six warehouses. The entire project took about a year.

"We spent a lot of time in the pilot phase, working out the bugs and learning the system," Timmons says. "It took several weeks to get through our first plant, and it became a smoother process as we went along to our other facilities."

During the early stages, it became obvious that the speech recognition used in a warehouse environment is nothing like that used in a customer contact center. Conveyors, fans, pallet jacks, forklifts, trucks, and people generate a lot of background noises that challenge the recognition software. Speech systems also have to contend with the many languages, accents, and nonstandard speech patterns used by the employees.

Eby-Brown employees are also using the system in the refrigerated and frozen sections of the warehouses, where there are loud compressors and a potential for condensation and ice, which can affect the microphones and mobile devices.

On top of all that, "each plant has some unique configurations," Timmons relates. "Essentially, it's the same system at each location, with some special site considerations built in," he says.

The Jennifer system, and the support staff from Lucas Systems, worked through each of these challenges flawlessly. "In the early stages of the project, we met with a lot of providers, and we liked Lucas right away," Timmons recalls. "They assured me that we wouldn't lose any of our customer service attributes and that we could tweak their system to accommodate our special needs."

Getting the employees to use and accept the system was a non-issue, according to Timmons. "Voice is so intuitive," he says. "A new employee can exchange data with Jennifer all night right from the start. The learning curve is so small."

Usually all that's required is a few minutes to train the system to accept each worker's voice. Then the worker needs only to log in at the start of his shift to begin using the system. It even stores his unique preferences and attributes.

The rollout of the Jennifer system preceded Eby-Brown's migration from a mainframe to an enterprise resource planning system and a switch from a legacy WMS to the much newer SAP WM solution.

"By the time we started our SAP WM rollout, we had already earned a full payback on the voice system," Timmons says. "With Jennifer in place, the transition to SAP was a nonevent for our pickers."

Jennifer's voice-directed technology uses open, standards-based interfaces to link into systems like the SAP WM. Lucas's building-block approach and service-oriented architecture meant that Jennifer could be tailored to Eby-Brown's unique needs in several weeks and its workers could be up and running quickly. "The system has already paid for itself," Timmons concludes.

App At a Glance

Since installing the Jennifer VoicePlus system in its seven warehouses, Eby-Brown has seen:

  • a 46 percent drop in the number of picking errors;
  • productivity, measured in the number of items picked per person in an hour, increase 33.4 percent for full cases and 40.1 percent for individual items; and
  • an increase in shipping volumes without making additional capital investments.

Page1 of 1