A unified platform to test, protect & govern AI at scale

SplxAI, founded in 2023 in New York, developed one of the more advanced automated security platforms for AI systems prior to its acquisition by Zscaler in late 2025. The company focused on securing the full AI lifecycle with particular strength in addressing risks introduced by agentic AI workflows. These included prompt injection vulnerabilities, hallucination-driven actions, unauthorized data extraction, and unsafe tool interactions.

At the core of SplxAI’s platform was an automated red teaming engine designed to simulate sophisticated adversarial scenarios. Its multi-modal threat detection capabilities analyzed text, images, and voice interactions for malicious patterns, while its runtime

guardrails enforced policy compliance directly within AI agents and orchestration environments. This positioned SplxAI as one of the first vendors to treat AI systems not as static models to protect, but as dynamic, autonomous actors requiring continuous controls.

The company’s “purple teaming” philosophy — intentionally blending offensive and defensive AI Security techniques — allowed organizations to accelerate vulnerability discovery and shorten remediation cycles. By reducing reliance on manual testing and continuously probing AI decision-making pathways, SplxAI helped enterprises identify emerging weaknesses before production deployment.

With a total of $9 million raised before its acquisition and strong early traction, SplxAI became an attractive strategic buy for Zscaler. This buyout marked one of the first notable acquisitions in the AI Security market and signaled the increasing importance of safeguarding agentic AI systems at scale.