Lasso AI Security Platform

Lasso Security is an AI security company founded in 2023 by Elad Schulman, Ophir Dror, Lior Ziv and Yuval Abadi, with its press releases datelined Tel Aviv and New York. It addresses the risks created when employees use generative AI tools and when enterprises build their own LLM applications and AI agents: prompt injection, data leakage, unsafe tool calls and unmanaged shadow AI.

The platform follows a discover, assess and protect loop. Discovery builds an AI bill of materials of agents, models, system prompts, tools and guardrails; AI security posture management finds misconfigurations and policy gaps; automated red teaming runs multi-turn and tool-chain attacks; runtime enforcement applies policies inline at the proxy, API or AI gateway layer; and detection and response maps threats to MITRE and OWASP. In 2026 the company added LEAP, a transformer-free guardrail that runs on ordinary CPUs, paired with RAPID, a self-hosted LLM-as-a-judge. Customers shown include the US Department of Homeland Security, eToro and Fiverr.

• Discovery and AI-BOM: Inventories agents, apps, models, prompts, tools and guardrails.
• AI security posture management: Finds misconfigurations, policy gaps and supply chain risk.
• Automated AI red teaming: Runs multi-turn, context poisoning and tool chain attacks from a large attack library.
• Runtime enforcement: Applies policies inline at the proxy, API or AI gateway layer.
• AI detection and response: Detects threats mapped to MITRE and OWASP with attack context.
• AI usage control: Governs employee use of GenAI tools, coding assistants and MCP servers.

A Vietnamese bank can put runtime guardrails in front of its customer-facing LLM chatbot to block prompt injection and leakage of account data. A software company can discover which coding assistants and MCP servers its developers use and set policies for them. A fintech can red-team its AI agent before launch and keep detections in its SOC.

Security teams get one inventory of every AI agent, model and tool in use instead of scattered spreadsheets. Inline policies block prompt injection and sensitive data leaks with a stated 98.6% detection accuracy and classification under 50 ms. The vendor claims its guardrails cost 570 times less than cloud-native alternatives, which matters at high AI traffic volumes. Red-team findings feed back into runtime policies, shortening the fix cycle.

Enterprises are moving from chatbots to autonomous agents that call tools and access internal data, which widens the attack surface beyond what traditional DLP and web firewalls cover. Analysts now track AI usage control and AI runtime defense as distinct categories. Banks and digital businesses in Vietnam and Southeast Asia that deploy GenAI assistants face the same need for guardrails and audit trails.

Lasso Security Inc. was founded in 2023. In September 2026 it announced a 30 million USD round led by ClearSky with Entree Capital, iAngels, Singtel Innov8, Mindset and Swish Data, and reported revenue growth above 500% over 12 months. It states SOC 2 Type II certification and was named a Gartner Cool Vendor in 2024.