Lasso Security

Lasso Security is a GenAI-first, end-to-end AI security platform that protects models, agents, and applications across the enterprise. The platform runs a five-pillar lifecycle. Discover autonomously inventories every AI model, agent, and application in the organization, including shadow AI tools employees adopt without IT approval, with platform integrations and CI connections that automatically surface homegrown applications. Assess analyzes security posture by mapping what each agent is exposed to, what actions it can perform, and what sensitive data it can reach. Test runs automated red teaming through offensive AI agents drawing on a library of more than 3,000 attack types and techniques across the OWASP Top 10, including multi-turn agentic attacks tailored to the customer's use case, with purple teaming that auto-updates policies from findings. Enforce applies policies inline at the proxy, API, or AI gateway layer. Protect detects and terminates threats at runtime mapped to MITRE and OWASP, delivering full context on every attack: what happened, which agent was targeted, the impact, and how to resolve it. In 2026 Lasso introduced the Intent Deputy, pioneering what it calls Intent Security for agentic AI: behavioral baselines that monitor not just what agents do but what they intend to do, extending coverage to MCP servers and prompt injection in agent workflows. The platform deploys via Gateway, API, or SDK with one-line integration, feeds SIEM, SOAR, ticketing, and messaging systems, integrates with Cloudflare for network-level protection, and is available in Azure Marketplace. Lasso serves financial services, healthcare, and regulated industries with HIPAA and GDPR compliance support, and formed Lasso Federal LLC for the US public sector. Alongside Noma and Zenity it is one of the remaining independent platforms in an AI security cluster that has consolidated rapidly into incumbents, differentiated by its guardrails-layer speed claims (sub-50ms classifications, a published 570x cost-efficiency claim versus cloud-native guardrails) and its intent-based approach to agent security.