Knostic [1] [2] [3] [4], a provider of need-to-know access controls for Generative AI [1] [2] [3] [4], has recently emerged from stealth and is a finalist in the 2024 RSA Conference Launch Pad competition [1] [2] [4].


With the rise of Large Language Models (LLMs) like Microsoft Copilot and Glean [2], organizations face increased risks of exposing sensitive information. Knostic’s innovative solution aims to enable mass adoption of LLMs for search and chatbot use cases by implementing knowledge-level controls based on business context [3]. This approach ensures that employees receive personalized answers tailored to their specific need-to-know boundaries [3], preventing LLM oversharing and potential business harm [3]. Founded by cybersecurity experts Gadi Evron and Sounil Yu [1] [2] [4], Knostic has secured $3.3M in pre-seed financing and has garnered support from investors such as Shield Capital and DNX Ventures. The company’s unique approach to need-to-know personalization is seen as a key enabler for enterprises and the DoD to leverage LLMs for AI-powered innovation while maintaining security and safety. Co-founders of Knostic relied on system thinking and mental models to develop this innovative solution for AI security and safety [3].


Knostic’s solution has the potential to significantly impact the way organizations utilize LLMs, ensuring data security and preventing potential risks. With the support of investors and recognition in competitions like the RSA Conference Launch Pad, Knostic is poised to make a lasting impact on the field of AI security and safety.