What Is VoxIQ ?
VoxIQ is an enabling technology that combines speech technology with databases, using Knowledge Based System techniques to decide the limited set of keywords that the speech recognition engine must recognise at any one time. This makes the task simpler for the speech technology which results in a significant improvement in speech recognition.
What does it actually do ?
A VoxIQ equipped system listens to a conversation and, by detecting key words; it establishes and maintains knowledge of the context of the conversation as it evolves. From this knowledge, the system can access existing application databases and display relevant information to allow the user to talk in greater depth on the subject and to ensure that the information displayed and Keywords being sought are always relevant.
Speech / database and KBS / database interaction already exists. Melding KBS with SR produces a unique benefit because the KBS can dynamically determine, in real-time, the relatively small subset of keywords to be listened for at any one time by the speech technology. The accuracy and speed of the recognition process is therefore dramatically improved, firstly in speed by reducing the number of words being searched for at any one time, secondly in accuracy by reducing the problems posed for conventional SR by dialects, accents or the fact that words with different meanings can sound the same. Thus the context of the conversation can be accurately determined and key words identified within that context, in real-time.
The applications developer can exploit the KBS technology to enhance the quality and relevance of information presented to the operator by the computer system. This can be done by embedding expert knowledge (i.e. the rules under which the system is designed to work) in the VoxIQ interaction database specifically set up for that application.
How is it used?
Take a simple example of VoxIQ being used by an insurance broker. When a customer calls and begins a conversation with the agent, the mention of such key words as car insurance and Ford Mondeo will cause the system to display appropriate forms on the agent's computer screen and it will populate these with existing customer and product information retrieved from the database(s). Further words such as disability or racing might be used. These would also be recognised and further information would be retrieved to prompt the agent with information regarding important ancillary conditions of the insurance.
What are the implications?
VoxIQ makes the use of speech technology highly efficient and leads to significant productivity improvements in a number of business applications, e.g. call centres, where it will lead to an increase in call throughput giving management the choice of expanding services or reducing the cost base. At the same time staff will be providing an enhanced level of service to customers.
Application software developers will also benefit from the speed and flexibility of VoxIQ as a development tool. It will be designed so that they are able to make simple additions and modifications to it without the need for the further involvement of VoxIQ Ltd.