Voice-recognition technology hit the warehouse scene nearly 15 years ago with major installations in the grocery and food industry, and companies such as Wal-Mart, Kroger, and Kellogg's have been successfully using it in distribution centers ever since. Yet even a few years ago, many warehouse operators still regarded this technology as experimental and esoteric. The idea of workers giving and receiving voice commands as they selected orders and replenished stock seemed more like 2001: A Space Odyssey than 2001: A Warehouse Reality.

That perception has started to change, and organizations that were on the fence about the technology five or six years ago are now showing serious interest, says Aaron Miller, principal with the consulting firm Tompkins Associates.

Voice technology vendors agree with that assessment. Steve Gerrard, vice president of marketing for Princeton, N.J.-based Voxware, says trade show attendees who used to come by his company's booth and say, "That looks interesting, but I don't know if we would ever have a need for it," are now asking how to integrate it into their operations. "The market has definitely gotten over a critical hurdle," he says.

One look at the numbers confirms their observations. The annual growth rate for Vocollect, for example, has exceeded 60 percent in each of the last four years, and the Pittsburgh-based company anticipates its annual revenue will leap from $40 million last year to $75 million in 2005. Competitor Voxware's revenue rose 91 percent in just one year, from $4.5 million in 2002 to $8.6 million in 2003, and the company is projecting revenues in excess of $11 million for 2004. And there's plenty of room for expansion: Gerrard estimates that just one percent of U.S. warehouses are currently using voice-recognition technology.

Why such rapid growth now for a technology that's been around for 15 years? Much of it is being fueled by vendors expanding their reach into new markets. Not only are they selling to other parts of the globe, but they also are seeing domestic sales take off. "We've seen an explosion into verticals outside of our traditional markets, including mass merchandising, specialty retailing, third-party logistics, consumer goods, and automotive aftermarket," says Tim Eusterman, Vocollect's vice president of marketing and business development. Technology providers, moreover, are adapting their products for uses such as cycle counting, receiving, and replenishment.

A third factor fueling voice systems' popularity is improved technology. "Four or five years ago, a lot of companies that I talked to that were looking at or experimenting with voice recognition just didn't think the technology was mature enough," says Steve Banker, service director, supply chain management at ARC Advisory Group. "That has changed in the last year."

According to Banker, the software has become better at reading workers' voices and screening out the noise of roaring freezer fans, rumbling conveyor belts, and slamming dock doors. It's also become more flexible, so the software can be tailored to the way a company chooses to process its product, instead of the other way around.

But the biggest attraction for potential users may be a fast payback—a critical point considering that voice technology doesn't come cheap. Prices vary by type of implementation and order volume, but a ballpark figure would be about $4,000–$5,000 per operator, says Marc Wulfraat, managing partner of consulting firm KOM International. By comparison, radio frequency/bar-code scanning runs $2,500–$3,000 per person. (That's in addition to overhead for the RF network infrastructure, software, support, and service.)

Despite that initial cost, companies usually see a return on that investment within the first year. "We started to get a payback right away," says Ken Collins, project manager for St. Albert, Alberta-based Connect Logistics, a third-party logistics provider. "But it was six months before we had all the training and processes in place where we could track a significant improvement and prove it to management."

What makes it possible to achieve ROI so quickly? With voice technology, order pickers aren't laden with paper pick lists, labels, or scanners. Nor are they shifting their attention between a paper list or tiny computer screen and the items they're picking.

As a result, both order accuracy and productivity rise. According to Eusterman, it's not unusual for companies to see 50 percent reductions in errors. Some users see even greater improvements: Wholesale food distributor U.S. Foodservice, for example, reduced picking errors by 70 percent when it rolled out Voxware's VoiceLogistics solution in 24 of its distribution centers.

For high-volume operations, even small improvements in accuracy can produce a substantial payback. Miller reports that Associated Wholesale Grocers' Kansas City warehouse improved its order-accuracy rate from 99.52 percent to 99.64 percent after implementing Vocollect's Talkman solution. At 62 million cases a year, that translated to an additional 74,000 cases picked correctly and a $1.5 million savings, he says.

More accurate picking can create a ripple effect that benefits other areas. U.S. Foodservice found that increased accuracy reduced off-day deliveries and returns while increasing customer satisfaction. "We track the number of errors, and we show large improvements there," says Jack Granahan, the company's director of voice deployment. "However, those aren't the true gains," he adds. "The true gains come from saving miles, reducing fuel costs, and customers getting their orders on time by reducing the number of off-day deliveries."

Voice technology has also boosted productivity for 99-Cent Only Stores at its Texas distribution center, and that's translated into big savings and a quick ROI, says DC Manager Don Arter. The fact that voice-directed picking is easy to learn helped to speed that payback. "The training is just unbelievably simple," says Arter. "What we like about it is, if we have someone who is lift-certified, we can have them trained on voice and out on the floor picking on the first day."

In some cases, using voice-recognition technology is a "no brainer." For example, it's ideal in cold environments, where employees must wear gloves that make it difficult to use handhelds. Another application is full-case picking of heavy items. When pickers are lifting 30-pound boxes for eight hours, having both hands free makes good ergonomic sense.

But no technology is appropriate for every company and every situation, and voice technology should be thought of as just another tool in the paperless picking toolbox, says Wulfraat.

Voice vendors are now moving into broken-case and individual-item picking, where the technology choice can be complicated. "Once you open up the brown box and start dealing with individual SKUs, you need to look at what the most productive and accurate technology is going to be," says Wulfraat. "Is it RF scanners? Is it voice technology? Is it pick-to-light? The answer is no longer clear."

Wulfraat recommends comparing the cost for each of those technologies to the projected savings. "Companies need to ask, why pay the extra capital for voice? What's the added advantage if hands-free is not an issue?"

Even if hands-free picking is needed, voice technology isn't the only way to go. A pick-to-light system, for example, might be more appropriate for high-speed picking. "The human eye can see a light faster than it takes for the brain to interpret a voice command," observes Wulfraat.

In some cases, voice technology may not be accurate enough. Voice systems typically use short location IDs instead of lengthy item numbers because they're easier to understand. If an item has been misplaced in the wrong bin or slot, however, the operator will unknowingly select an incorrect item.

When it does make economic sense to implement a voice-recognition system, though, it's likely to pay for itself by improving order-picking accuracy, productivity, ergonomics, and customer satisfaction—and do it in record time. That's why for more and more companies today, voice-recognition technology is no longer the stuff of science fiction but has instead become a financially rewarding reality.