Speech Technology Magazine

 

Speech Gives the Orders

A medical supplies distributor improves order accuracy with a voice-directed logistics system.
By Leonard Klie - Posted Apr 2, 2009
Page1 of 1
Bookmark and Share

Henry Schein, a distributor of medical, dental, and veterinary supplies, has a standing policy that all orders received by 5 p.m. are to be processed and shipped the same day. But with five distribution centers scattered around the country shipping an inventory of more than 190,000 stock-keeping units (SKUs)—90,000 in stock and another 100,000 available as special orders—to roughly 250,000 small private practices and clinics throughout the United States and Canada, it’s a constant challenge to process client orders and get them out the door as quickly and accurately as possible.

To ensure fast and accurate delivery of orders, the Melville, N.Y.-based company installed voice-directed logistics technology at its warehouses. The first facility, in Jacksonville, Fla., went live in August 2006, followed by Dallas in November 2007, Denver in October 2008, and Reno, Nev., in December 2008. Its facility in Indianapolis just completed its rollout in January.

However, for Mike Charpin, senior project manager at Henry Schein, interest in voice technology had been building for some time prior to the first installation. “This project had its origins in 2001, but at the time voice was primarily only in the grocery industry, and it was still very cost-prohibitive,” he says. “At the time, I did think that voice was something we should consider going forward, though.”

His interest in voice revived in 2003, when Charpin says he decided to start “seriously” looking at specific vendors, their applications, and the potential return on investment.

After what he calls “a pretty exhaustive search,” Charpin selected the Jennifer solution from Wexford, Pa.-based Lucas Systems. Incorporating speech recognition and text-to-speech capabilities, Jennifer uses information from a host or warehouse management system (WMS) to hold two-way conversations with warehouse workers throughout their shifts. Taking cues from the WMS, Jennifer instructs workers on what to do and where to go throughout the facility, assigns items to be picked from the shelves for pending orders, and responds to and acts on the information workers send back to the system. Users interact with the system by speaking into their headsets, typing on the keypads on their portable computers, or scanning information using the integrated barcode or radio-frequency readers. 

Prior to voice, if a worker was sent to a particular location to pick three items and found only two there, he had to stop what he was doing, fill out a form, and bring it to the office. Someone from the office then had to physically go to the location to verify the claim and direct someone on the floor to replenish the supply.

Now, notification of exception conditions, such as product stock-outs or damaged goods, can be made to a supervisor in real time. “Using voice, we can do it all automatically at the point of the pick,” Charpin says. “We’ve definitely streamlined that process with the voice application.”

And while productivity gains are common when voice technologies are used in the warehouse, Charpin says his company was more interested in improving picking accuracy. “The real benefit is in accuracy and the integrity of the orders we send to our customers,” he states. “As a business, as a company, Henry Schein has always been good at managing same-day shipments, and our shipping accuracy was always very good. And even though we’ve had very high accuracy, we’re always looking to raise it.”

And that’s just what voice has achieved. Early in the deployment cycle, product-picking accuracy increased by 12 percent, primarily because the system collects verbal confirmation from the worker for every action he takes. At the same time, Jennifer lets managers pinpoint the date, time, location, item, and quantity of each action and easily trace any error to its source.

Validation Required

In addition, Jennifer can validate the quantity of each order with a countdown feature to assist on high unit count activities. Jennifer also allows workers to ask the system to repeat a task order or provide additional product information, such as lot number, product code data, or expiration dates, if clarification is needed. 

Other benefits have been substantial, as well. Productivity, measured by the number of product lines each worker can process per hour, improved by 8 percent. Training time for new workers has been cut from weeks to days because Jennifer offers multilevel user modes— from learning to expert—to help new employees get up to speed quickly. 

A new user can get additional, more detailed prompts that are different from those given to more experienced users. Jennifer also features verbal controls that allow users to adjust the speed of voice prompts on the fly, based on their personal preferences. Jennifer can speak more common or frequently used prompts faster than those used less frequently.

Overall, the technology has been so promising that Henry Schein is looking to expand its use later this year. Currently, only 250 workers at the five distribution centers use Jennifer, and those workers use it only for picking loose items, which make up about 46 percent of the 60,000 orders per day that leave Henry Schein’s warehouses.

“If you order a full case of a product, we’re not using voice for that,” Charpin says. But that is likely to change. “We’re looking to use voice for case picking. We see it as a viable application to pursue later this year to further streamline things,” Charpin says.

Besides the expected ROI numbers, Charpin was also attracted to Lucas’ Jennifer solution because of its open-source, service-oriented architecture. Each Jennifer solution is assembled using a library of standard software building blocks that incorporate customer-specific business rules and application logic. 

“We wanted an open platform because our business already was working well, and we did not want to have to change how we do things,” Charpin says. “The Jennifer system is not a basic application that was modified for our business model, but was written by us and for us with Lucas.”

“The interface was probably the easiest part of the installation. It was all straightforward and really went pretty smoothly,” Charpin maintains.

The biggest hurdle during the implementation was more of a personal—or personnel—issue than a technological one. “The installation went well, but we had a high concern for our folks and how they would view [the technology] and the use of headsets,” Charpin recalls. It turned out to be less of an issue than originally expected.

“The Jacksonville team was excited about being the first [to use voice], but there was also a lot of apprehension toward the unknown, some Big Brother kinds of concerns,” he says. “We designed the system so it wouldn’t be intrusive, and we spent a lot of time in Jacksonville making sure we understood [the workers’] apprehension and addressing their concerns.”

Charpin also spent time at each location explaining to workers what the voice application would entail and what the company expected to achieve by employing it. “Once Jacksonville and Dallas went up, the unknowns became knowns and fears dropped across the company,” he says. “We took the time to get things right for ourselves and our people. Once we got a little momentum going and word spread from one [distribution center] to the next, momentum built and things went pretty quickly.” 

Charpin and his colleagues also expected language to be an issue because Henry Schein selected the English-only application, even though many of the company’s workers speak a lot of other languages. “But so far we haven’t had anyone who was not able to use [the system],” Charpin says.


App At a Glance

Since installing the Jennifer voice-directed logistics system from Lucas Systems in its warehouses, Henry Schein has seen:

  • product-picking accuracy increase by 12 percent;
  • productivity improve by 8 percent; 
  • supervisors gain visibility into work in progress; and 
  • training time for new workers cut from weeks to days.

Page1 of 1