Know-how is the most important asset of a high-tech company. Typically, information is concealed in technical specifications, use and maintenance handbooks, patents of invention, contracts, procedures, standards, regulations etc. In high technological content markets, mastering knowledge and making it available at the right time is crucial to gain competitive advantage.
We propose a brand-new RPA (Robotic Process Automation) approach, based on automatic text mining from company documents, in order to retrieve valuable information by means of NLP (Natural Language Processing) algorithms. Furthermore, retrieved information can be organized in knowledge structures tailored on customer’s needs.
Our solutions enable:
Valorisation of existing know-how by structuring and making available the information
Relieving users from repetitive, time-consuming and non-value-added tasks
Saving time and costs on documentation analysis
Enhancing objectivity and repeatability of processes
Augmentation of company knowledge, by drawing information from external sources
TagMyText is an entity recognition software module for automatic information retrieval in the internal company documentation. Information retrieval is made possible by means of custom-specific knowledge bases, tailored on client’s needs. The module provides an augmented readibility interface and is the starting point for the advanced modules (Classifier and Matchmaking).
Classifier is a software module that allows the accurate and flexible classification of company documentation based on the textual content. After a short training phase aimed to generate the classification categories that best map a documentary dataset, the software tool allows the automatic classification, organization and navigation of the obtained database, boasting a very high degree of reliability.
Matchmaking is a software module for the creation of networks between unrelated documents on the basis of their textual content. By detecting terminological relationships, Matchmaking is able to structure a documental dataset in conceptual graphs or hierarchical trees, recreating the natural knowledge distribution. In this way, it is possible to take over both inflexibility of relational databases and low accessibility of non-structured data.