A Discussion with VISO TRUST: Expanding Atlas Vector Search to Provide Better-Informed Risk Decisions

VISO TRUST, an AI-powered cyber risk and trust platform, has made significant advancements in its use of MongoDB since September 2023, notably adopting MongoDB's new Search Nodes architecture to improve search scalability and efficiency. They have enhanced their information extraction techniques from PDF and image files, enabling more accurate retrieval augmented generation (RAG) systems that rely on dense and sparse embeddings for efficient data extraction and retrieval. These improvements have bolstered their intelligent question-answering system, which utilizes MongoDB Atlas Vector Search for quick and accurate responses to security inquiries by parsing and answering questions based on uploaded compliance documents. VISO TRUST's technological evolution reflects a broader trend of integrating large language models (LLMs) and discriminative model inference to optimize data handling and risk assessment processes, with MongoDB playing a crucial role in storing and querying the vast amounts of training data required for these operations.