Vanta signs deal for Miami-linked AI security startup Pillar
The agreement would connect Pillar's agent testing and protection tools with Vanta's governance platform, with new functions planned in phases during 2027.
Key Takeaways
- •Investor Golden Ventures lists Pillar's location as Miami and names founders Dor Sarig and Ziv Karliner.
- •Vanta plans AI inventory and monitoring for January 2027, followed by testing and enforcement later that year.
- •Pillar focuses on discovering, testing and protecting AI systems that access data and take business actions.
Vanta signed a definitive agreement Nov. 9 to acquire Pillar Security, bringing a Miami-linked startup's AI protection technology into its security and compliance platform. The companies did not disclose financial terms in their announcements.
Pillar appears in Golden Ventures' portfolio with Miami listed as its location. The investor identifies Dor Sarig and Ziv Karliner as its leaders, connecting the transaction to South Florida's technology community.
Sarig and Karliner founded Pillar in 2023. In their joint announcement, they described a business built for AI systems that do more than generate answers: agents can use tools, reach company data and take actions for an organization.
Connecting policies to software
Pillar says its platform finds AI applications, tests them for weaknesses and monitors their behavior. Its protection tools are designed to stop actions that fall outside an organization's rules while the software is running.
The founders describe business context as part of that work. For an agent handling customer records, security teams need to understand its purpose, permitted activity and who is responsible for the risk. That information helps determine which tests to run and where to set limits.
Vanta plans to combine those capabilities with its Trust Graph, which connects a company's security policies, controls and responsibilities. The aim is to make an AI-use policy enforceable in the software carrying out the work, with evidence showing when protections operate.
A staged rollout in 2027
Vanta's roadmap starts in January 2027 with an inventory of AI applications, agents, models and connected tools, alongside monitoring for suspicious activity. The company says those functions would help organizations understand what is operating and where to investigate.
Testing through simulated attacks is planned for spring 2027. Runtime enforcement, intended to block unauthorized actions, is scheduled for the summer. Vanta says its paid AI Governance offering will become available in January.
Pillar's founders frame the combination around two connected questions for businesses adopting AI: what a system should be allowed to do, and whether its behavior stays inside those boundaries. Their proposed answer brings testing and protection into the same program as governance.
