Telephonic communications – Audio message storage – retrieval – or synthesis – Voice activation or recognition
Reexamination Certificate
2002-06-13
2004-05-04
Hoosain, Allan (Department: 2645)
Telephonic communications
Audio message storage, retrieval, or synthesis
Voice activation or recognition
C379S088070, C379S088130, C379S088160, C379S265010, C379S265090, C379S266010, C379S266070
Reexamination Certificate
active
06731722
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the field of automated transaction processing and more specifically to a system and method for processing transactions that minimizes the use of humans, or live operators, while retaining many of the benefits of live operator processing.
BACKGROUND OF THE INVENTION
The field of direct marketing encompasses diverse transactions. These transactions often occur over a telephone, but are increasingly being conducted over nontraditional media including the internet. Types of transactions include processing dealer location inquiries, answering consumer product inquiries, taking customer orders, and troubleshooting product problems. Ideally, live operators would process all transactions. Live operators offer many advantages over automated systems. For example, live operators are able to recognize and deal with problem transactions, such as calls by angry customers. However, the cost of using live operators to process transactions is often prohibitive. As might be expected, when the operators are used exclusively, labor costs constitute a large percentage of the total cost of processing telephone transactions.
The high cost of using live operators has led to the use of automated systems for processing these transactions, especially transactions conducted over the telephone. However, known automated systems suffer from important drawbacks that make their use undesirable for many applications. Such systems are difficult and/or time consuming to use, often requiring users to listen to multiple menus and make choices using a touch tone telephone keypad. As a result, customers frequently become frustrated with known automated systems.
These drawbacks have led to systems that combine automated and human operators to varying degrees. Some of these partially automated systems use automation to route a call to a particular live operator. An example of such a system is one used for processing customer calls to a company for troubleshooting one of that company's products, wherein the automated portion consists of prompting the customer to select the problem product from a list of the company's products and connecting the customer to a live operator who can help the customer.
Systems such as these, however, are only partially effective at reducing costs for several reasons: First, the amount of automation achieved is relatively small. Second, the initial, automated portion of the transaction may be objectionable to some callers. Third, many inquiries may be more complicated than the automated choices. What is needed is an automated transaction processing system that automates as much of a transaction as possible without alerting the user that the system is automated, or an automated transaction processing system that automates a transaction in such a way that the user will not object to having the transaction processed by an automated system.
One labor intensive task that cannot be performed by known automated transaction processing systems is the generation of mailing addresses in computer recognizable form from address information spoken by a user of the system. Ideally, a speech recognition system would be used to perform this task. However, while present day speech recognition systems have the ability to recognize limited vocabularies from speaker independent sources, or recognize large vocabularies from speakers whom the speech recognition system has been trained to understand, they cannot yet recognize large vocabularies from speaker independent sources over the public switched telephone network. The size of a vocabulary needed to encompass all of the geographic and proper names in even a relatively small geographic area far exceeds the size of the vocabularies of the systems available today. Therefore, using speech recognition systems for automatically generating name and address information is not yet practicable.
One method for automatically generating name and address information in a computer recognizable form which is practicable today is to use the Dial Number Identification Service, or DNIS, (which provides the telephone number of the calling telephone) and ANI (automatic number identification) commercial address databases to generate an address based on the number of the phone used by the caller. This method is not very reliable, however. This is because the best commercial databases are only approximately 70% accurate, because calls are sometimes made from phones other than a caller's home phone, and because a caller may not be the member of the household listed in the database records even if the caller is calling from home, such as one spouse calling in an order on a phone listed only in the other spouse's name.
As a result, human operators are required to convert spoken name and address information to a computer recognizable form. The best systems today use a combination of human operators and ANI/commercial database methods. Human operators compare the database address obtained via ANI with the spoken response (either interactively or using recorded calls) and make any necessary address corrections to the database addresses. These methods, while an improvement over completely manual methods, are still labor intensive because a human operator must listen to every call to determine whether the database address matches the spoken address. What is needed is a system that can automatically recognize and indicate whether spoken name and address information matches name and address information in computer recognizable form.
Alternative communications media are also becoming increasingly important marketing tools. For example, the internet is becoming an increasingly acceptable alternative to telemarketing. Other technologies involving televisions and combinations of television and internet loom on the horizon. Thus, today's automated transaction processing systems must be designed to operate with these alternative media.
SUMMARY OF INVENTION
The present invention provides an apparatus and method for automating customer service transactions. The apparatus comprises a plurality of automated devices, such as voice recorders and players, voice recognition units, and audiotext address generators, as well as a plurality of live operator terminals. The invention treats a transaction such as a telephone call as a series of subtransactions. Each subtransaction has data associated with it. The data may be supplied by the system, such as data corresponding to a prompt to be made to the user; or the data may be supplied by the user, such as the user's response to a system prompt. Automated devices and live operators are treated as resources for processing the data. A transaction is processed by switching the data associated with each subprocesses to one or more appropriate resources.
Efficient use of human labor is made for two reasons. First, live operators are generally only used to process critical subtransactions or subtransactions that cannot be processed by automated resources. Automated resources are used to process the remaining subtransactions. This includes the use of voice players to play previously recorded predetermined prompts and messages, and the use of automated resources to process user input where possible and desirable. As will be discussed in further detail below, the decision as to whether to use live operators for a particular subtransaction is based on cost and quality considerations. Second, resources, including live operators, are only assigned to a particular transaction while they are required for a subtransaction. When a resource is not in use for one particular transaction, it is available as a resource to process other transactions.
The system is controlled by application control blocks that are developed for a particular application. The application control block acts as a script by defining the series of subtransactions necessary to process a particular transaction and defining the content of the prompts and messages that will be communicated to a user by the system. The applic
Callfx.Com
Dilworth Paxson LLP
Hoosain Allan
McConathy Evelyn H.
Shorter Darryl W.
LandOfFree
Automated transaction processing system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automated transaction processing system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automated transaction processing system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3258520