WitnessAI

WitnessAI has developed a platform to enable the secure and effective adoption of enterprise AI, particularly large language models (LLMs) and generative AI. It ensures data privacy and regulatory compliance by using isolated instances encrypted with customer-specific keys. It intercepts activity between users and LLMs, applies control policies, and protects data and user safety with millisecond latency, without requiring installation on users' devices.

The WitnessAI Secure AI Enablement Platform offers three main capabilities:

Witness/OBSERVE: Provides visibility into employee AI usage, allowing IT departments to monitor which AI systems are being accessed and how they are being utilized.

Witness/CONTROL: Enforces AI-specific policies and governance, such as preventing LLM training data from being used for unauthorized purposes, restricting access to customer data within a private LLM, or blocking company IP from being shared with public LLMs like ChatGPT.

Witness/PROTECT: Secures enterprises from potential data or financial loss due to AI misuse or attacks, addressing risks like prompt injection, jailbreaking, and hallucination.

WitnessAI is deployed as an isolated, cloud-based instance for each customer, encrypted with their unique keys to ensure data privacy and regulatory separation. The platform intercepts activity between users and LLMs, applies control policies, and ensures data protection and user safety, all with minimal latency and without requiring installation on users' devices.

Funding and Financials

In May 2024, WitnessAI raised USD 27.5 million in a Series A funding round co-led by GV and Ballistic Ventures. Funds were earmarked to fast-track the development of more AI guardrails and to build a global sales and support organization.