Lasso
Discover. Attack. Defend. Continuously. We break your AI before someone else does - then we stop them.
About
Lasso secures AI wherever it runs inside a company: employees using AI tools in daily work, coding agents writing and shipping production software, and the custom AI applications engineering teams build and deploy. One continuous loop discovers that activity, assesses its risk, red-teams it with autonomous and operator-steered attacks, and enforces guardrails inline at runtime - with every red-team finding feeding back into the guardrails automatically. What makes that loop practical at scale is the architecture underneath it. A patent-pending CPU-only classification engine inspects prompts, responses and tool calls in single-digit milliseconds at effectively zero marginal cost, so protection stays on across 100% of production traffic rather than being sampled or switched off under load. Lasso's intent engine goes beyond content filtering: it detects when an agent's actions diverge from what the user actually asked for. Lasso deploys as SaaS, in your VPC, on-premises or fully air-gapped, and maps every application to NIST AI RMF, OWASP LLM and Agentic Top 10, MITRE ATLAS and the EU AI Act. Named in Gartner's Market Overview for AI Usage Control (2026) alongside Microsoft, Cisco, Zscaler, Palo Alto Networks and Check Point, and to the inaugural Allied Defense 100 in the Cyber vertical. Certified to ISO 27001, SOC 2, GDPR and PCI-DSS, with ISO 42001 completing in Q4 2026. In production at Fortune 500 enterprises and US federal agencies, including air-gapped environments.
Related Vendors
HiddenLayer provides an AI security platform that secures agentic, generative, and predictive AI applications across the lifecycle. The platform unifies supply chain security, runtime defense, posture management, and automated red teaming to protect models and agents. It aims to prevent prompt injection, data leakage, model tampering, and agent abuse while protecting intellectual property and supporting compliance needs. HiddenLayer offers automated agent and model scanning that integrates into CI/CD pipelines for transparent, automated scanning of models and agent artifacts. The platform also provides runtime defense by integrating with existing security infrastructure to deliver visibility, detection, and response for AI applications at enterprise scale.
2021.AI offers the GRACE Governance platform, which ensures comprehensive AI Governance, AI Risk Management, and Compliance Assurance according to both external regulations and internal policies. The GRACE LLM 2.0 facilitates efficient operation of large language models (LLMs) with enhanced safety features and real-time monitoring. It includes tools to manage sensitive data, enforce policies, and support compliance with evolving regulations. With an on-premise option, it gives organizations full control over data security, meeting strict compliance requirements. Integrated coding assistants further enhance productivity while maintaining code privacy.
Innodata is a global data engineering company with over 35 years of experience, serving leading technology companies. The company specializes in AI initiatives, including industry-leading generative and traditional AI solutions. Their offerings encompass supervised fine-tuning, red teaming, data collection and annotation, and AI consulting. Innodata focuses on developing and implementing custom strategies tailored to clients' specific needs, helping enterprises drive AI innovation. Their commitment to AI TRiSM (Trust, Risk & Security Management) ensures secure and reliable AI implementations that are easy to monitor.
