The hottest technology improves, the price drops,

  • Detail

Technology improvement, price decline, voice recognition is new.

voice recognition system sometimes makes consumers "angry". For example, this system often prompts "please say your customer number", but no matter what the user says, the system cannot recognize it. However, the voice recognition system is still attractive when the company's decision makers make business plans, because the cost of customer service will be greatly reduced after using this system. Even if the call center is transferred to the third world countries, the cost reduction can not be compared with the benefits brought by the use of voice recognition system

due to the new development of technology, the speech recognition system has become easier to use, and began to enter the enterprise market slowly, with slight progress in sales. This technology has appeared more than ten years ago. It is mainly used to replace the audio key service system (to query the voice service information through the digital keys)

improve the work efficiency of customer service

the above two systems can reduce the workload of customer service representatives. This kind of position must deal with the increasing customer service requests, which requires a lot of repetitive work, and these two systems can replace the customer service representatives to complete simple and repetitive work. If the voice recognition system is combined with the "text/voice" or "voice/text" conversion system, the whole customer service process can be automated. As enterprises pay more and more attention to customer service, speech recognition system begins to attract more and more attention

speech recognition system has three traditional shortcomings: low recognition rate, need of special equipment and high price. System developers have been trying to solve these problems over the years, especially in improving the recognition rate. "With the improvement of processor operation speed and program algorithm, the recognition rate has been greatly improved," said Steve cramoysan, chief analyst of Gartner. More and more speech recognition systems can even recognize some common dialect sounds

questions about user questions and answers

art Schoeller, senior analyst of Yankee market analysis company, said that even if the system can correctly identify what users say, new problems will come again

when the customer calls the voice recognition system for the hydraulic jaw of Jinan gold testing machine, his speech content may be very complex, and the voice recognition system cannot effectively handle it, so the customer's speech content must be limited. In other words, let the voice recognition system ask closed-ended questions to customers, such as "are you using your home phone?". The customer's answer to this question is limited and easy to handle. The voice system does not ask open-ended questions, such as "what do you want us to do?" The user may answer too many questions, which will be difficult for the system to handle

to limit the feedback content of customers, enterprises must spend a lot of time to design voice recognition programs, sort out internal business processes, determine the problems that the system can handle, and then integrate them into the IT business department if the industry continues to adopt the mainstream solvent based process to produce PU synthetic leather

the final result is that enterprises using speech recognition systems must pay a lot of fees to suppliers to design and deploy speech recognition programs. For every dollar spent on speech recognition systems, it will cost $6 on services and supporting facilities

in order to simplify the deployment of voice recognition system, developers are designing templates, such as those specially used for financial business. With regard to templates, it is not possible to determine how much they are worth to users

standard problem

there is another problem left over from history: system design. Developers always provide their own proprietary equipment, and users will have great difficulties in deploying and maintaining the system. In addition, users should not only purchase the manufacturer's special equipment, but also purchase the corresponding voice recognition program development tools

the voice XML (VXML) standard has been published in its second edition. Its function is to make use of the advantages of XML development in speech recognition. VXML can simplify program development and has a very positive impact on the cultivation of speech recognition market

standards are valuable, but not a panacea. Developers will further develop according to the standards, and each improvement will increase the difficulty of integration between products of different suppliers. Microsoft's move also complicates the issue of standards. Microsoft does not fully follow VXML, but is promoting another standard, salt (voice programming language markup), which provides voice extensions for a variety of markup languages (including HTML and XML). Cisco and Intel are firmly on Microsoft's side, but most traditional speech recognition developers (such as IBM, nuance and scansoft) do not pay much attention to Microsoft's standards

as the developers gradually start to adopt the standard to cut a lot of 500mm from the first end of each coil of steel wire and then take two samples, the product development will be simplified and the price will be reduced so far. In fact, the price of speech recognition products and services decreased by about 30% in 2004

gradually becoming popular

with the improvement of technology and the decline of price, voice recognition products will be favored by more enterprises. Speech recognition technology has been common in many industries, such as aviation, finance and telecommunications. Recently, it has also been applied in other fields, such as publishing, hardware and software support. In addition, some companies have begun to use voice recognition systems in their internal communications. For example, employees who go out to troubleshoot can clear the fault declaration through this system after completing their tasks

the development of voice recognition system is slow and stable, and no rapid and exciting progress has been made. One of the reasons is that many companies have adopted web-based customer service systems, which are easier to implement than voice weaving systems. Therefore, the future of speech recognition is bright, but not plain sailing. Experts predict that enterprises will gradually use voice recognition system to replace the existing audio button service system, and the market will grow by 5% ~ 10% every year. (end)

Copyright © 2011 JIN SHI