Business Intelligence and Customer Insight in the Enterprise is improving fast because of two important phenomena.
The techniques for natural language processing, advanced analytics and machine learning are becoming almost a commodity through the hard work of companies like Amazon, Google, and the Data Science Community.
More and more data in the enterprise has become available for analytics.
We now routinely take unstructured text in the Enterprise and use natural language processing (NLP) techniques to extract values and features from written natural language.
But within the area of unstructured information, often termed “Dark Data”, there are still opportunities where we can expect a lot of gain. One key area of Dark Data development is in the domain of spoken interactions between company representatives and customers.
Currently companies have few options other than to record conversations much less perform analytics on the content of these calls. Yes: IBM, Google and Amazon offer speech recognition that theoretically could be used to turn voice into text and then use NLP techniques, but the problem is that these systems are not able to be “trained” on Enterprise specific taxonomies.
The accuracy and value to the Enterprise is contingent on these systems being able to manage industry specific lingo, product names, and acronyms to name a few requirements.
During this presentation we will demonstrate an approach for taxonomy driven speech recognition so that industry specific spoken text can be adequately recognized for NLP processing and subsequent use in a customer Knowledge Graph.