What WitnessAI does
WitnessAI is the enablement platform for safe enterprise AI use, sitting between users (and AI agents) and large language models to enforce policy, data protection, and observability in real time. The product intercepts prompts and responses to public chatbots (ChatGPT, Claude, Gemini, Copilot), enterprise copilots, and internally built AI agents, then applies controls — block, redact, route, or log — based on user, group, data classification, and intent. WitnessAI runs in milliseconds, with nothing installed on user devices, making it a network-level guardrail that works across the entire AI surface area an enterprise exposes.
In 2025-2026 the company expanded into agentic AI governance, giving security teams visibility into autonomous agent activity — what tools agents call, what data they touch, and how they behave over time — plus the ability to set policy on agent runtime behavior, not just on chat prompts. WitnessAI reports over 500% ARR growth and 5x headcount growth in the past 12 months, with 73 employees and a fast-growing roster of Fortune 500 customers.
Who it's for
WitnessAI targets CISOs, security architects, and AI governance leaders at large enterprises rolling out GenAI to thousands of employees and dozens of AI agents, who need a single control plane for safe use across public and private models.
Pricing
WitnessAI is sold via custom enterprise pricing tied to seats, AI traffic volume, and add-ons such as agent governance and red-team services.
Team & funding
WitnessAI was founded in 2023 by Rick Caccia (CEO, ex-Symantec, Palo Alto Networks, Exabeam) and Gil Spencer (Co-founder). The company was incubated by Ballistic Ventures and headquartered in California. WitnessAI announced a $27.5M Series A in May 2024 co-led by GV (Google Ventures) and Ballistic Ventures, followed by a $58M strategic round in January 2026 led by Sound Ventures with Fin Capital, Qualcomm Ventures, Samsung Ventures, and Forgepoint Capital — bringing total funding to $85.5M.
Position vs competitors
WitnessAI competes with Lakera, HiddenLayer, Pillar Security, Prompt Security, Robust Intelligence (Cisco), CalypsoAI (F5), and Lasso Security. Its differentiation is network-level inline enforcement plus agentic governance, with no agent installed on end-user devices.