The JUMP project aims at bringing together the knowledge stored in different information systems in order to satisfy information and training needs in knowledge-intensive organisations. Electronic Performance Support Systems provide help, advices, demonstrations, or any other informative support that a user needs to the accomplishment of job tasks in her day-to-day working environment. The paper describes the JUMP framework, which is designed to offer multiple ways for the user to query the knowledge base resulting from integration of autonomous legacy systems. Semantic Web languages and technologies are used throughout the framework to represent, exchange and query the knowledge, while Natural Language Processing Techniques are implemented to understand natural language queries formulated by the user and provide consistent and satisfying results.