The growing maturity of speech recognition technology, coupled with the constant improvements in the price and performance of high-end processors, has presented companies with a viable alternative to proprietary DTMF (dual-tone multi-frequency)-driven applications.
That`s the word from Shaun Cochrane, GM of Intelleca, the local representative of the Vocalis range of speech recognition products.
"The business case for speech recognition technology is built around lowering operational costs, increasing revenues and providing higher levels of customer service," says Cochrane.
"Speech recognition technology can be delivered via an onsite and application service provider (ASP) model, making it affordable even for smaller businesses."
Speech recognition systems address many of the shortcomings of traditional DTMF systems, which depend on the audio signals generated when a user presses the buttons on a touch-tone telephone.
Speech recognition eliminates the need for touch-tone responses, allowing customers to interact with the system using voice - a natural interface that allows for hands-free and eyes-free operation and faster completion of transactions. In addition, speech recognition can deliver advanced applications such as travel reservations, sharing trading and banking transactions that are not practical with traditional DTMF systems. For example with speech recognition, a user can issue a command such as "Please transfer R200 from my cheque account to my credit card next week Friday", which the system will recognise.
The technology can lower operational costs and improve customer service by shortening the total length of a call through reducing hold times and transfers. Mundane and repetitive calls can be forwarded to the speech recognition system, allowing call centre agents to focus on complex and high-value customer interactions.
Speech recognition technology can be used to classify call types more efficiently and to gather information before the call is transferred to an agent or even to complete the entire call without agent intervention, saving time and increasing productivity in the contact centre.
Cochrane says speech recognition technology can either be integrated with an existing call centre infrastructure or built on a flexible Internet-based architecture. Using the Internet as a platform means that new applications can be developed and deployed quickly and easily using HTML code, eliminating the need for specialist skills.
It is also a simple matter for companies to extend services from their Web sites to the telephone using speech technology. All they need do is script shadow Web pages in the HTML language, and users will be able to connect to the website through a gateway.
Web site data, encoded in standard HTML, is accessed and retrieved, and supplied to the caller after it is converted by the text-to-speech technology. The user`s spoken command is converted back into a format the system can understand through a similar process.
"Because this class of speech recognition is built on an open, Internet architecture, it represents the next generation of call centre technology. It provides a platform on which companies can build solutions that pull together their Web, mobile and call centre strategies," says Cochrane.
Cochrane says speech recognition technology breaks speech down into a set of algorithms used to decode the user`s speech patterns into a form the system can understand. Rapid improvements in processing power in line with Moore`s Law are allowing speech recognition systems to handle algorithms of ever-increasing complexity, which means systems can deliver more sophisticated functionality and accurate results.
For example, speech recognition systems now include features such as:
* Barge-in technology that allows callers to interrupt the application at any time;
* Natural language processing and word-spotting capabilities that allow users to speak in complete sentences and phrases;
* Speaker-independent recognition which ensures that the application can recognise different languages and dialects;
* Confidence-checking technology that allows the system to automatically question the caller until it gets a sensible response if it is not confident of the results it obtained.
The technology has a wide range of potential applications - ranging from call routing and interactive banking services, through to delivering flight information, share prices, and sports results over the telephone.
Share