Speech Technology Magazine

 

Around the World with Voice Response

The challenge of attracting, retaining and developing relationships increases as companies expand to world markets. Each country has unique sets of technical, cultural and communications issues to overcome. Customer relationship management is an ongoing continuous improvement process being used worldwide and customers are demanding premier service delivery.
By Kelly M. Lumpkin - Posted Jun 30, 1997
Page1 of 1
Bookmark and Share
The challenge of attracting, retaining and developing relationships increases as companies expand to world markets. Each country has unique sets of technical, cultural and communications issues to overcome. Customer relationship management is an ongoing continuous improvement process being used worldwide and customers are demanding premier service delivery.

Most companies have goals to deliver premier service, as a strategic necessity. At the same time, these companies must be able to deliver service at a cost they can afford. Teaming automation with human resources is key to cost-effective, international, around the clock service. Companies are investing in voice response systems (VRS) with specialized applications to meet customer demands. As VRS applications are deployed worldwide, they must be "internationalized."

A VRS uses recorded messages to prompt a caller for input. Callers provide input by pressing keys on touch-tone telephone keypads to generate Dual Tone Multi-Frequency (DTMF) tones. The VRS collects the DTMF tones and converts them into the data value as indicated on the telephone face plate, and in turn sends the data collected to a host computer or computer program for processing. After the data is processed, the VRS can take the resulting return codes, convert them to speech by concatenating recorded message and play the response back to the caller over the telephone.

The banking industry is a pioneer in the use of VRS. They have addressed repetitive applications such as routing telephone calls, looking up account balances, transactions histories, loan payments, answering requests for fax documentation and generally conveying routine, dynamic information. Without VRS companies are limited to using human resources to perform these functions replete with international staffing, training, operations and management challenges. VRS can be easily operated 24 hours a day, seven days a week, at a fraction of the cost of live service representatives.

Challenges

VRS deployment outside North America has not been as rapid due to the following obstacles:

  • Rotary or pulse type telephone sets
  • Phones do not have letters on the face plate
  • Requirements for multi-lingual speech recognition.

    Rotary or pulse type telephone sets can not generate DTMF tones that a VRS uses for collecting data. The absence of alphabet characters on the telephone face plate means callers can not be prompted for alphabetic input. Temporary fixes include Pulse-to-Tone converters (PTTC) and portable tone generators.

    Neither of these devices gained wide acceptance because they have lacked reliability, speed of operation and ease-of-use. Vendors seemed to lose sight of what they were asking users to do. Imagine a caller holding a portable tone generator to the mouthpiece of a telephone, listening to prompts, pressing buttons on the mouthpiece and at the same time reading a bank statement.

    Speech recognition allows the telephone set to be used as it was intended. Callers place a call, speak into the mouthpiece and listen to the receiver. Even better, information can be accessed while the callers hands and eyes are free to do other things.

    Multilingual speech recognition technologies are available in over 25 languages and dialects. This means that the VRS can now reach areas around the world. And the cost of the systems and application development suggests that the cost of answering a telephone line with a sophisticated multilingual speech enabled application is less than $1.00 per hour. The cost advantage over human resource labor costs can now expand to most of the developing world.

    To be commercially practical, the VRS application should be designed with a Voice User Interface (VUI) for caller interactive applications. VUI applications can ignore the limitations of a telephone keypad and be designed to be faster, easier and more natural to use. For example, with a typical VRS, callers are asked to press the number of the key that is associated with the action they want. Most designers limit prompting to no more than four choices per menu to make the choices easy to remember. With a speech enabled VUI, a caller will speak the action they want and go directly to it. There are no associations to remember!

    Financial Applications

    Imagine a financial services application with multiple functions including locating local branches in your city, state, province or other general area. If the country market area contains 50 sub-divisions like the states in the US, a very complex set of time consuming menus would be needed.

    Let's consider a function similar to locating a state in the United States. With a standard VRS design, 13 menus of 4 choices are needed to specify one of the 50 United States. Each menu would take about 15 seconds to read. If the state of interest is Utah the caller would have to listen for over 200 seconds to hear the prompt for their state.

    With speech recognition enabled VUI, they would just speak the name of the required location and could reduce the menu prompt to less than five seconds. The same principle can be applied to other sub-divisions and is particularly valuable in international applications where callers can not spell words on the keypad.

    Speech recognition capability also allows callers to spell out words. If they are interested in a specific credit card transaction, they can spell the name of the transaction in question. The VRS can quickly search the available transactions and locate the appropriate record.

    Increasing Security

    VRS applications in industries such as financial services have stringent needs for security. The present standard for controlling access to confidential applications is a 4-digit PIN (personal identification number). These numbers are initially provided by the service provider and are frequently forgotten. With speech recognition, we use the caller's spoken input for speaker verification.

    The caller may be prompted to speak a password. The system will then analyze the utterance for word match against a previously enrolled database. At the same time, it will verify the utterance against the voice characteristics in the database to help insure the caller is the person they claim to be.

    With a 4-digit PIN, there are only 9,999 possible combinations of digits. With a VUI with speaker recognition, we can include letters, words, or phrases. The words and phrases are part of a speaker dependent recognition function and will therefore be easier to remember than a 4-digit random PIN. At the same time, the caller generated spoken password offers an infinite number of possible combinations. The service provider will also benefit by dramatically few calls to their help center to re-establish or reset forgotten PINs.

    Building a Business Case

    Multilingual speech recognition offers compelling competitive advantages for companies expanding customer service delivery in international markets. Among the reasons:

  • Cost advantage over human resources for routine functions.
  • Reduced telephony costs due to shorter call duration
  • Increased security using VUI speaker verification
  • Deploy systems in markets with high rotary telephone penetration
  • Expand into non-English markets
  • 24/7 Operation is possible. (24 Hours a Day, 7 Days a Week)

    In summary, international multilingual speech recognition and speaker verification technology can provide easier-to-use and more secure VRS applications. If a forward thinking company goes into a market with applications in the local language that are designed to be efficient and easy to use, that market will respond favorably to their products and services. What's more, service providers will reduce operations costs and free resources to help attract, retain and develop their most profitable customers.

    Kelly M. Lumpkin is the director of business development for Alternate Access, Inc., a computer telephone integration company specializing in speech recognition, text-to-speech conversion and other VRS and call center applications. For more information, visit the web site http://alternateaccess.com or send Email inquiries to info@alternate access.com or call (919) 781-8371. Caller Interactive Applications is a service mark of Alternate Access.

  • Page1 of 1