The objective of this research activity is to design a software module designed to index texts from different sources. The knowledge base creation process starts with crawling: this is the phase that extracts content, interpreting different formats (html, txt, pdf, doc/docx). The next phase is processing, where the semantic index by scanning the unstructured data within the documents and identifying the relationships between the concepts they contain. The resulting knowledge base supports Customer Care and can be queried through keywords or phrases in natural language, integrated through APIs in various applications (CRM, chat-bot, voice bots and more).