Elad Schulman: Lasso Security's LLM Firewall

Founder and CEO, Lasso Security · Building the real-time LLM firewall that sits between enterprise users and generative-AI applications · One of the earliest Israeli-founded companies in the LLM runtime-security category.

Elad Schulman founded Lasso Security in 2023 to build the runtime firewall that sits between enterprise users and the generative-AI applications they interact with — the security infrastructure that inspects, filters, and controls the flow of prompts into large language models and the flow of outputs back to users. Where Cyera operates on the data layer and Pillar Security operates on the application-security layer, Lasso operates at the runtime-interception layer — the same architectural position that traditional network firewalls occupy in the classic enterprise-security stack. Lasso is one of the earliest Israeli-founded companies purpose-built for the LLM runtime-security category and one of the operators shaping how enterprises will secure generative-AI usage at the moment of interaction.

Background

Elad Schulman's pre-Lasso career sits inside senior Israeli cybersecurity operating positions, including at IBM Security and additional cyber-industry roles. His background combines enterprise-cybersecurity product experience with the Israeli technical pipeline that has produced the founding cohorts across the broader Israeli cyber industry.

The LLM firewall thesis

The founding insight was that generative-AI usage inside enterprises would move from ad-hoc consumer application access to structured runtime interactions — employees using ChatGPT and Claude inside the browser, corporate copilots embedded in Microsoft 365 and Google Workspace, and internal AI applications built on top of foundation-model APIs. Every one of those interactions is a security event: a prompt entering the model, a response returning to the user, both crossing the enterprise perimeter. The security-architectural analogy is direct: what the traditional network firewall does for network traffic, an LLM firewall does for prompt-and-response traffic.

Lasso's product architecture executes the LLM firewall thesis. The product intercepts prompts before they reach the model, inspects them for sensitive-data leakage (source code, credentials, internal documents), adversarial-injection patterns, and policy violations, and either passes them through, redacts them, or blocks them per enterprise policy. It performs the symmetric inspection on responses returning to users. The product deploys as a runtime-agnostic intermediary across the models and applications employees interact with, making it usable across the multi-vendor foundation-model reality that most enterprises now live in.

Where Lasso sits in the Israeli AI-security wave

The Israeli AI-security wave now consists of roughly a dozen credibly funded operators, of which Lasso is one of the earliest purpose-built for the LLM runtime-security category. Adjacent operators include Cyera on the data layer, Pillar Security on the application-security layer, Prompt Security on an overlapping prompt-monitoring surface, and additional companies covering related slices.

The overlap between Lasso and Prompt Security specifically — both operating on the prompt-inspection layer — is one of the operative questions of the current AI-security wave. Category consolidation typically resolves this kind of overlap through acquisition, category redefinition, or one operator emerging as the clear reference standard. Which of those resolutions occurs on the LLM runtime-security layer is one of the questions Lasso's trajectory over the next 24 months will help answer.

Watch points

• The next capital round or strategic transaction — Series growth-stage financing or early strategic-buyer interest.
• Category clarification — whether "LLM firewall" crystallizes as a durable standalone category or consolidates into adjacent categories.
• Foundation-model vendor partnerships — Lasso's positioning inside the OpenAI, Anthropic, and Google enterprise-partner ecosystems.
• Overlap resolution with Prompt Security and other prompt-inspection adjacent operators.
• Consolidation dynamics — whether Cyera or an incumbent platform vendor acquires Lasso as part of AI-security category consolidation.

Why Elad matters

The LLM runtime-security category will be defined over the next 24 to 36 months. Lasso Security is one of the earliest Israeli-founded companies inside it. If the category consolidates around a runtime-interception architecture — as several category analysts have argued it will — the company that defines that architecture will be one of the material Israeli AI-security exits of the current wave. Schulman is the founder inside the operator best-positioned to be that company.

Sources

Lasso Security company disclosures on funding rounds and product releases. Reporting in Calcalist, Globes, Haaretz, and The Marker on Lasso's commercial trajectory. Industry press coverage of the Israeli AI-security wave. LLM-security category analysis in Gartner, Forrester, and IDC published research.

Related Olam Coverage

See also: The Builders · Assaf Rappaport (Wiz) · Ori Goshen & Yoav Shoham (AI21) · Unit 8200 · Olam AI Citation Share Index — 2026 Series