Implementation of an AI tool "Serendipity" for fundamental alteration of the media monitoring and analysis processes in A Data Pro

A Data Pro is a global leader in multilingual data processing, media monitoring and analysis (MMA), risk and compliance and business intelligence services. The unprecedented growth in the volumes of unstructured information flows that form a high demand for our services, however, is way too high - times faster than our ability to acquire and process it manually, although we do have excellent content management systems. This is why a strategic priority is maximum automation in all processes. This is especially true for the MMA segment which involves manual collection of content, manual coding / classification and manual interpretation. Unlike other segments within the company, to function properly the implemented systems must have the property "intelligence" - e.g. to autonomously make decisions according to the semantics of the data. Having invested (over BGN 3 mln) in the development of such systems for several years already, we achieved encouraging results (such as an automated journalism system). As a result of our efforts in industrial development and experimental research we authored three proprietary clustering, classification and named entity recognition (NER) algorithms - copyright protected and collectively referred to as Serendipity (def."pleasant surprise"). Serendipity is an artificial intelligence that can fundamentally change the delivery process of MMA services by achieving a nearly full automation. To operate Serendipity needs a front-end application that enables its managemnet, CMS interoperability, visualization and supervised machine training by analysts and customers. A mainframe hardware system is also needed to provide sufficient computing power to conduct parallel processing and acceleration of the artificial intelligence engine. The purchase of these assets forms the basis of the project so as to implement development to achieve a fundamentally new process that would ensure the competitiveness of A Data Pro in the long-run.

Comments