How does FlexAnswer Work?

FlexAnswer utilises state of the art natural language and information classification technologies. The advanced natural language processing capabilities of FlexAnswer's underlying technology allows the service to give precise answers to questions; an approach which is simply not possible with traditional search engines. Classification technologies allow the system to achieve a high level of performance in the cases where the advanced natural language technology can not be applied. A powerful learning mechanism allows the system to benefit from previous questions and their answers so that it can automatically improve over time.

The combination of these technologies can be a powerful asset to any company that wishes to enhance its customer support activities. The FlexAnswer service allows a company to benefit from these technologies, and to deploy them in a timely and cost effective manner.

Natural Language Processing

FlexAnswer utilizes next generation natural language processing capabilities to address the problems of understanding and responding to customer queries. The technology on which FlexAnswer has been built has been developed over nearly 15 years in both academic and commercial environments.

While many products claim to use natural language technology, they typically do so in a superficial way. Approximately only half of the question answering solutions today on the market attempt even spell correction; very few of these attempt to tackle any of the other problems of natural language analysis. Our natural language processor is able to take into account: spelling mistakes, morphological form, grammatical structure, semantic, pragmatic and dialogue information.

The knowledge that the system uses includes a vocabulary of over 160,000 word meanings and is organised in a hierarchy so as to take advantage of the relationships between different words and meanings. In addition to word related information the system also contains large amounts of knowledge related to how words can be used in different situations and how these can combine together to generate various interpretations of a question.

By using its full range of natural language processing techniques, together with its extensive knowledge sources, our technology is able to generate a 'deep' logical representation to capture the meaning of a question. This allows FlexAnswer to give precise answers to questions; a benefit that is not available to other purely statistical or 'surface' approaches.