Voice picking technology in the warehouse is a proven and highly successful technology. It is simple to use, liked by operatives, delivers 99.99 per cent picking accuracy rates and gives return on investment in just months rather than years.

However, some adopters of voice have chosen unproven voice solutions and failed to get value for money and regretted their investment in voice. Indeed a number of major organisations have been driven to change their voice solution due to an initial poor technology choice.

But it is not the voice recognition solution itself that is to blame but the use of inappropriate hardware for the task in question and apparently low-budget, ‘one size fits all’ solutions that can be the root of the problem. Traditional radio frequency (RF) suppliers are under considerable pressure to avoid losing market share and are adding voice to their range of services. But in some cases the devices chosen may be inappropriate to meet the customers needs.

Voice software can be fitted and function on any Windows CE device, whether it is a dedicated wearable voice computer, a rugged handheld PDA or even a mobile phone. The question does not lie with the quality of voice software but the capabilities of the hardware on which it is installed.

Voice technology is proven to give best results in a rapid picking operation when used on a ‘hands-free, eyes-free’ device linked to a headset. It is the quality of the audio which is key; it is vital that the device can recognise the operative’s voice first time, every time, regardless of language or accent. The difference between picking on a voice unit and picking from a piece of paper might only be seconds per pick, even fractions of a second per pick depending on the operation being performed. So if the operative has difficulty using the unit and they need to repeat themselves you can very quickly lose those small portions of time.

Voice picking on a handheld device has been around for a long time and with certain devices it is proven to be successful. But worryingly in a number of circumstances voice has been implemented on inappropriate hand-held devices delivering wholly unsatisfactory results. There is a market for voice on a handheld device. In a warehouse that is not fully voice-driven and multi-modal interaction is necessary, a voice-enabled hand-held device can add real benefits. For example, one task in the warehouse requires voice picking but for another a screen display is needed, or another task requires a built-in scanner.

However, not all hand-held devices are adept at handling the audio requirements of high quality voice recognition in a rapid pick, high noise environment such as a warehouse. But voice on a hand-held was never originally intended to be used for high volume, high density picking in a warehouse environment and in many instances deploying a one-size fits all hand-held RF solution will deliver sub-standard voice performance with poor accuracy and recognition of voice and slow speed of response. Talk to the people who have tried it!

Compare it with a modern mobile phone which can perform a number of functions above and beyond its basic purpose - to make and receive phone calls. The 3G Internet connection can download web content in seconds, the two million pixel camera can take high quality digital photos and Bluetooth connectivity can allow the user to instantly share those images with other phones for free. But for all these marvels, if the phone gives a poor or unreliable connection and you can’t make and receive calls when you want to, then all additional functions are pointless. Because you can’t do what you set out to do - make phone calls.

Furthermore, multi-modal functionality in a warehouse is increasingly less necessary; voice technology fully integrated into a number of warehouse management systems is capable of handling all warehouse requirements from goods in, to goods out.

Many people believe cost will be one driving factor. While the cost of a hand-held device will typically be less than a hands-free, eyes-free device, the hand-held device will not include the software required to run the voice application. Once that cost is added in the price differential should be comparatively small. While initially this may be a cost saving, in the longer term return on investment will actually be slower. If the audio capabilities of the hand-held device are not good enough order pickers will have to repeat themselves more often and productivity will be lower, not to mention a higher level of damage to the hardware because hand-held terminals are proven to be less robust and more susceptible to breakages than a highly robust dedicated voice unit.

The desire to buy a future-proof solution is another factor. Warehouse managers are keen to make a wise investment for today and tomorrow. In some cases they will buy a solution that performs functions they think they may need in the future, so they may decide that they may want to perform multi-modal functions in the future and therefore believe that voice on a hand-held gives more flexibility for non-voice uses of the hand-held. This is not always prudent. Business managers should consider what their core requirements are now and buy a solution that satisfies those specific needs. Compromising on today’s definite needs in the hope of meeting the possible needs of tomorrow is typically not the best decision.

Voice is increasingly being seen as a ‘must have’ technology for the warehouse replacing hand-held terminals. Vocollect, whose Talkman unit is the world leader in warehouse voice applications, announced in June the 10,000th Talkman user in the UK, continues to show clear evidence of the rapidly increasing acceptance of voice technology. As a result the traditional RF suppliers are fighting hard not to lose their share of the market to voice and are starting to push voice on hand-helds.

In doing so some suppliers may be carrying out questionable certification procedures. I am increasingly sceptical about the number of hand-held devices which have been certified to work with voice. Any Windows CE device can run the software but I do not believe that the majority have been tested to a high enough standard within a working warehouse environment. They may not be able to filter background noise, cope with changes in the operatives voice level or give an audio output to the user that is loud enough in a noisy warehouse environment. Loose and faulty phone jacks on the device are also a common complaint causing crackling and feedback which may be disastrous to normal operations.

The recent decision by a major voice provider to certify a wide range of hand held devices by a range of suppliers is one that has certainly raised eyebrows within the industry. It is very hard to believe that all devices will have the level of audio capability to deliver an acceptable level of quality. It is a broad brush statement to make if they have not tested the range properly and the only true test of that is in the warehouse.

Ultimately it is the warehouses and distribution centres who could be losing out. A number of companies have had to acknowledge that they do need a dedicated voice terminal and that voice on certain hand-helds has not worked. Voice on a hand-held does most certainly work where the audio capabilities of the unit are up to the job, but when purchasing a voice solution warehouse managers must make tough decisions.

It is tempting if they have always been tied to a particular RF supplier for all their RF requirements to just go for that supplier’s voice-enabled hand-held solution, but they should seriously investigate if the provider has access to hardware suitable for their voice requirements. They must carefully consider what their specific needs are and whether implementing a ‘low-budget’ solution will actually solve all the picking needs. Failure to do so will result in a long, expensive and ultimately unsuccessful implementation.

BIOS, Oct 05, 06 | Print | Send | Comments (0) | Posted In Audio