WitnessAI Raises $58M to Build ‘Confidence Layer’ for Enterprise AI Security

New updates have been reported about WitnessAI.

WitnessAI has secured $58 million in new funding to develop what it describes as a “confidence layer for enterprise AI,” targeting the growing risk that corporate use of AI chatbots, copilots, and autonomous agents could expose sensitive data or breach compliance rules. Speaking on TechCrunch’s Equity podcast, CEO Rick Caccia and investor Barmak Meftah of Ballistic Ventures outlined how enterprises are increasingly concerned about prompt injection attacks, inadvertent data leakage, and uncontrolled interactions between AI agents, particularly as generative AI tools are embedded in workflows across functions. The funding underscores rising investor conviction that AI security will become a core infrastructure category as organizations scale AI from pilots to production.

Caccia and Meftah framed WitnessAI’s opportunity within a projected AI security market that could reach $800 billion to $1.2 trillion by 2031, driven by demand for governance, control, and observability over AI systems deployed in highly regulated and data-sensitive environments. WitnessAI aims to position itself as the intermediary layer that governs how employees and AI agents access and use powerful models, enforcing policies to prevent confidential information from escaping and to ensure regulatory compliance as AI autonomy increases. A key emerging risk they highlighted is AI agents interacting directly with other AI agents without human oversight, which could amplify errors, propagate policy violations, or create opaque decision chains. For executives, the implication is that AI adoption at scale will require a dedicated security and governance stack, and WitnessAI is building specifically to fill that role, seeking to become embedded in enterprise AI architectures as a strategic control point.