The pace of technology is faster than conventional security settings as businesses work to incorporate artificial intelligence into everything from script technology to consumer service. A business is betting that protection in the era of agentic, self-executing software may require a fundamental rethink, not just an update.
With$ 9 million in seed funding, Israeli startup Pillar Security has emerged from the shadows to address what it sees as a growing security gap in the enterprise AI stack. Shield Capital led the round, while Golden Ventures, Ground Up Ventures, and a group of corporate angel investors joined them.
Wall, which was founded in October 2023 by Dor Sarig and Ziv Karliner, aims to serve as the basic level of AI-native software, especially as businesses implement agentic devices capable of intelligent decision-making and conversation.
Artificial “radically alters the way we build software,” according to the statement from AI. It introduces a completely new lifecycle rather than just adding another step to conventional processes. said Dor Sarig, CEO and co-founder of Pillar Security. In the era of brains, software has agency and data is downloadable. With this knowledge in mind, Pillar’s systems, which is supported by real-world AI threat intelligence, creates a new class of protection specifically designed for security risks posed by AI. Application security is being redefined to resemble the intelligence-hegemony’s agentic and intelligent program.
Relevant reports:
Two in five organizations have already experienced AI-related security or private incidents, including unit adjustment, data leak, and Internet robbery, according to a new Deloitte survey of 1,200 cybersecurity leaders. The risk was destructive in 25 % of those situations.
Attacks that don’t always follow recognizable patterns, such as jailbreaking LLMs ( large language models ), prompt injections ( training sets ), or data poisoning (using outdated tools ), are unreliable and can’t be identified and remedied.
The goal of Pillar’s approach is to treat AI systems as powerful, risk-sensitive resources, mapping everything from the information and prompts that train them to the run-time system they depend on. From there, the platform creates adaptive guardrails, red-team simulations, and constantly assesses operational risk based on real danger telemetry. The company claims that it has much more detail to identify and avoid AI-targeted attacks than traditional security software.