What does Skimle do?
Skimle draws inspiration from social science methodology to leverage AI in an entirely new way to supercharge your knowledge work. Our iterative process turns your body of documents, such as interview transcripts, call audio files, group interview video tapes, reports, statements, data room presentations, or contracts into structured knowledge, helping you identify the insights that matter the most.
Think of Skimle as your tireless research assistant. You define what types of insight you want to extract from your document, or let the AI suggest based on the contents of the documents themselves. Insight types could follow your interview guide, your desired output topics (for example, market characteristics, competitors, key purchasing criteria and so in a Due Diligence) or more loosely specified criteria such as pro's and con's mentioned. When you upload the documents, the Skimle AI workflow processes your documents thoroughly to identify relevant insights and categorize them into a parsimonious and comprehensive category scheme from your data. The categories and subcategories essentially synthesize observations across your documents, helping you move from individual documents, such as what one interviewee told you, to more abstract questions, such as what all interviewees say about a specific issue.
Skimle's main view is the interactive "spreadsheet" that allows quick navigation from broad categories to individual insights and the content of documents. This Skimle table represents each document as a row, and each category as a column. This type of data structure is familiar and easy to work with for humans and allows zooming into individual categories (including the underlying quotes) as well as showing what all insights were extracted from each document. You can sort, filter, highlight, change columns and perform other intuitive tasks.
Clicking the magnifying class icon on any quote immediately takes you to the context of original document, helping you zoom into the raw materials. Even though AI is fairly good at analyzing and grouping content, the design philosophy of Skimle is always to keep the expert in the driver's seat!
Skimle features a highly customized chat interface that leverages the abstract knowledge structure and behind-the-scenes reasoning process to answer your queries. Our chat does not only draw on the original document materials like all widely available "RAG Chatbots" (Retrieval Augmented Generation), but also utilizes the insights and the categorization scheme to serve you a superior, structured response. The Skimle table also makes it easy and fast to export the insights into a structured report - be it a Word document showing what has been said about each theme, or a Powerpoint presentation with an executive summary and selected quotes per theme. Our unique Tree view allows you to dig deep into categories for analysis. We are working on API and MCP access as well to allow agents and other software to tap into the powerful structure provided by Skimle.