The Extract tool is built to systematise data from PDF documents with research, technical and scientific content. It extracts data from text, tables and some graphs and images, and links the values to the client's own desired output (an ODL, Output Data Layout). Data can be obtained in excel/csv files, JSON files or recorded directly in a database. No human made taxonomies or training is needed to set up the system. The system can achieve Precision and Recall of 94%/86%, which is better than human accuracy, and each extraction is fully automated and takes seconds, meaning time savings are immense.
The Researcher Workspace is build for Researchers in corporate R&D and academia to better handle vast amounts of research documentation. It is a content-centric platform that provides the user with a range of smart, AI-based tools to better navigate, review, filter down and extract data from research documents like papers, patents and internal documentation. 75% time savings have been shown, opening up researchers' time for more value creating tasks. Tools include a content based exploratory recommendation engine; an analysis tool for full document sets; smart filters based on either the machine's analysis or the user's own context descriptions; automatic summaries of multiple documents; and automatic extraction of table data. The platform allows the users to load any type of research documentation into these tools; from Open Access or Paywalled collections, patent collections, their own PDF collections or BibTex/CSV or similar exported files from reference managers.