Lasso’s Intent Deputy secures AI agents through real-time behavioral intent analysis

Lasso Security launched Intent Deputy, a behavioral intent framework designed to secure AI agents at runtime. It delivers real-time insight into AI behavior by interpreting intent, decision flow, and operational context.

“Intent Security represents the breakthrough security paradigm this rapidly evolving market demands, and Intent Deputy is our first-of-its-kind solution delivering it. It equips security teams with precise behavioral baselines to identify and stop agent deviations, caused by misconfiguration, behavioral drift, or malicious intent, in real time, before they escalate,” said Elad Schulman, CEO Lasso Security.

Enterprises are shifting from isolated chatbots to autonomous AI agents that reason, plan, and execute multi-step workflows across sensitive systems—accessing data, calling APIs, and acting independently on users’ behalf. This creates a non-deterministic attack surface where context and intent determine whether identical content is benign or malicious.

Legacy security tools relying on regex classifiers, keyword scanning, or high-latency LLM-only detection fail here. They miss sophisticated threats like indirect prompt injection, tool misuse, or excessive agency, while imposing performance penalties that degrade the end-user experience. Intent Security addresses this gap by shifting the focus from static content analysis to dynamic understanding of intent across the full interaction lifecycle.

The rise of tools like OpenClaw exemplifies the urgency. OpenClaw empowers users with significant autonomy, but misconfigurations or manipulations can expose credentials, local files, and connected services, or trigger rogue behaviors such as mass spamming.

“OpenClaw is shadow AI with a broader blast radius,” Schulman continued. “Employees adopt these tools because they genuinely help them work better, but the autonomous scale amplifies familiar vulnerabilities into enterprise-wide threats. Prohibition isn’t viable; the path forward is intelligent governance through deep behavioral visibility.”

Intent Deputy brings this new category to life as the runtime intelligence engine of the Lasso AI Security Platform. It moves beyond superficial, content-based scanning and instead analyzes the behavioral intent behind every AI interaction, delivering deeper and more meaningful protection.

Ophir Dror, Lasso’s CPO, explained, “Legacy tools only look at content: what the user asked or the AI output. Intent Deputy asks the real question: Is this action aligned with the agent’s authorized purpose? A marketing agent suddenly generating code? That’s often scope drift, not malice. By creating unique behavioral fingerprints for agents, as we do for humans, we detect anomalies in real time, deliver explainable audits for compliance, and let enterprises adopt agentic AI safely.”

Intent Deputy delivers on the Intent Security vision through four core capabilities:

• Intent alignment, not data patterns — Verifies that every action an AI agent takes is aligned with its authorized purpose by analyzing the full session history and context, not just scanning individual prompts for regex patterns and keywords. This catches threats that legacy tools miss, such as a customer service agent that begins exfiltrating data through a sequence of seemingly benign requests.
• Real-time protection without the performance tax — Delivers 99.83% threat detection at sub-50ms speeds, ensuring security never degrades the AI user experience. 570× more cost-effective than comparable cloud-native guardrail approaches.
• Behavioral baselines for users and agents — Creates unique “fingerprints” based on historical usage patterns, enabling detection of anomalous behavior even without malicious intent, the same approach enterprises already use for human identity security. ​​ When a user or agent’s behavioral fingerprint deviates from their established baseline, Intent Deputy flags the anomaly in real time.
• Explainable compliance — Provides human-readable evidence that AI operated within its authorized scope, giving legal and compliance teams the audit trail they need.

“As we scale AI agents across the enterprise, the security challenge isn’t just what data flows through them, it’s understanding why they’re taking the actions they take. Legacy tools may force enterprises to choose between security rigor and agent performance. Lasso’s approach brings behavioral visibility into an AI environment at the speed business requires, without compromising the user experience that drives adoption,” Itzik Menashe, CISO & VP IT Productivity at Telit Cinterion, concluded.