CUTTING EDGE TO LEADING EDGE: Voice Technology Emerges

Yes, it has emerged, and I’d like to take credit for the title, but that honor goes to Gerry Greenleaf, vice president of distribution for Hannaford Bros. Co., who is currently knee-deep in the implementation of a voice system. It appears that he’s right about the status of voice technology in logistics and distribution. He’ll also have more to say shortly.

Voice technology applications are increasing in popularity as a means to improve productivity in the warehouse. Voice, as the term is used here, comprises software applications in which a human operator (the user) interacts with some version of an automated aural communication system. In the course of this interaction, the human operator may receive instructions, provide input (feedback on location, quantity on hand, condition of the goods, or disposition), or facilitate information-gathering about performing a task (time or location). The system may respond by providing validation for the human input, alternatives for the operator to consider, or instructions for the next step in a process.

Communication can be either one-way — the system delivers instructions or information to the operator, who in turn executes the instructions — or two-way, in which the operator “talks back.” Until recently, the one-way version was probably more widely used, as it required fewer system resources. With the present state of processing hardware, that is no longer a significant factor. The personal hardware consists of a headset and a belt pack with the power supply and a wireless data-processing device (thick client).

Contemporary voice applications offer significant benefits and have resolved many of the problems of earlier versions (see sidebar on p. 13). Computer processing power has improved response time and reduced the hardware requirements. The training requirement to implement a voice system or to add a new user has been simplified. Most current warehouse applications will use a vocabulary of less than 100 words, and speaker-dependent recognition requires very little “system” training. Operators are said to get up to speed faster than they would by using RF devices.

Early buyers of voice were industry leaders who either were willing to take the initial risk of a new technology or could put enough weight and resources behind their choice to be fairly confident in attaining the results they sought. When the likes of Wal-Mart and Kroger sign up, credibility rises rapidly. Hannaford’s Greenleaf readily admits watching early adopters in the grocery industry work through their initial struggles to get to the proven product stage. In his own geography and industry, both Giant Eagle and Price Chopper preceded the Hannaford effort successfully. Now, with a pilot in one of his warehouse dairy zones recently completed, Greenleaf is ready to expand use of the application to the other major pick areas throughout his network.

However, like bar-code scanning and radio frequency before it, the wide adoption of a technology by the grocery industry has led to an expansion into other vertical markets. Dan Keller, director of sales and marketing for Lucas Systems ( and a veteran of voice systems installations, would add the liquor industry and retail and drugstore chains to the list of emerging markets. He expects food service to follow soon. If you talk to Vocollect (, arguably the most successful vendor offering voice solutions, based on its impressive growth in 2002, it would quickly add office products to the market list, and so on. In short, voice is coming to your neighborhood soon, if it’s not already there.

Vocollect cites improved accuracy and productivity as the two most important reasons for using voice. Keller of Lucas Systems believes that technological improvements, the credibility lent by big-name adopters, more rugged equipment, and device-independent applications account for the shift in buyers’ behavior. And, because of lower hardware costs, cost justification is becoming more favorable as well. Says Keller, “The ROI for almost every system we have bid on recently is less than one year.”

For Hannaford Bros. Co., pilot results showed significantly higher throughput, fewer selection errors and re-picks, less replenishment, and rapid implementation, with operators becoming as productive as before within five weeks.

The consensus in the industry is that voice has a bright future. Ivan Perez-Mendez, vice president of engineering at WMS provider Cadre Technologies (, anticipates that the entry price for access to the technology will soon drop considerably. Among the reasons he cites are the appearance of new providers, including Cadre, and the rapid shift to device-independent systems. Perez-Mendez also sees wireless pocket devices as playing a significant role in voice use. Lucas’ Keller does not predict as large a drop in pricing, but does concur that prices will be lower by the end of this year, and that there will be more players in the marketplace.

Meanwhile, established vendors continue to enhance their products, and users can soon expect more features. For example, most voice work thus far has concentrated on picking. Now vendors are developing templates for and refining generic rather than custom software for receiving, putaway, replenishment, and other functions. The integration between voice and WMS systems will also improve.

Your voice technology options are likely to proliferate quickly in the coming year, but some fundamental selection criteria pertain. The voice recognition engine should be robust, flexible, and sufficiently powerful to support a variety of uses and a large number of users. Response times should support your best users. Communications devices should be durable, including rugged construction and at least an eight- to ten-hour battery life. Training for both system and users should be simple and quick. Vendor support and development resources are essential, and as the technology evolves and matures, make sure that your operation has fast access to the improvements.

Before you apply voice technology, be clear on what it has to do in your environment. Then apply the technology intelligently. If voice enables communication or capture of data related to a poor method of accomplishing the work, it will be of no value and may even degrade performance. And finally, don’t lose sight of your ultimate goal — to do a better job.

Ron Hounsell is vice president at distribution consulting firm Tom Zosel Associates in Long Grove, IL. He can be reached at (847) 540-6543 or

Historically Speaking

Voice technology is not new, but distribution centers have been slow to adopt it. From its first appearance, voice appealed because it enabled an operator to use both hands and eyes on the task. However, this appeal was tempered by a number of shortcomings. First, because of limited computer processing power, response time was less than ideal. Second, the earliest versions of voice required extensive training associated with vocabulary and the individual operator’s voice and accent or dialect. Some applications also were affected by ambient noise in the work environment. Third, computer processing capacity inhibited response time. Finally, the cost per operator was high, especially in comparison to bar-code scanning, whether batch or radio frequency (RF). This meant that the technology could be justified only in a narrow range of situations. If inventory accuracy mattered most, operators could be precise and, by virtue of the hands-free capability, also efficient. The handling of high-value, serially numbered products such as electronics, therefore, lent itself to this kind of approach. Data capture or documentation for controlled substances or evidence-trail-related items are other examples. Some returned-goods operations found justifiable benefits, too. But for the average wholesale, retail, or fulfillment operation, the technology was not cost-effective.