Reducto: Making Human Data LLM-Ready With State-of-the-Art Accuracy

In this podcast episode, Diana Hu interviews Adit and Ronak, the founders of Reducto, a company that transforms complex documents like insurance claims and financial statements into structured data for AI applications. They discuss how they pivoted from their initial idea of long-term memory for LLMs to focus on document processing after realizing the struggles companies faced with data ingestion. Reducto's approach uses computer vision to parse documents like humans, improving LLM accuracy and enabling new use cases for their clients. They also share their experience of closing a deal with a Fortune 25 company and their efforts to create high-quality datasets for document extraction and are now expanding their team.