Artificial intelligence is rapidly becoming a staple of enterprise security operations, but many vendors still expect customers to trust proprietary AI agents hidden behind closed systems. Elisity is taking a different route.
The identity-based microsegmentation company has introduced the Elisity Command-Line Interface (CLI), an open interface that enables security and network teams to connect their own AI agents—including those powered by ChatGPT or Claude—to the Elisity platform. Rather than relying on vendor-managed AI automation, organizations can build custom agentic workflows while retaining full control over permissions, approvals, and security policies.
The release reflects a growing enterprise preference for transparent, auditable AI systems as organizations adopt AI-driven security operations without sacrificing governance.
An Open Alternative to Black-Box AI Security
Unlike many cybersecurity vendors that bundle proprietary AI assistants into their platforms, Elisity’s CLI gives customers a programmable interface to automate tasks using AI tools they already trust.
Teams can use the CLI to:
- Discover and inventory network assets
- Manage Zero Trust least-privilege access policies
- Generate posture and compliance reports
- Validate segmentation controls continuously
- Build custom AI-driven operational workflows
The CLI exposes 466 commands covering the Elisity Cloud Control Center API, while supporting output in JSON, YAML, CSV, and table formats for easy integration with enterprise automation platforms.
For organizations building AI-powered security operations, the platform also includes an operating guide and command glossary designed for large language models such as ChatGPT and Claude. The documentation helps AI agents execute valid commands instead of generating incorrect or unsupported ones—a common challenge when integrating LLMs into enterprise tooling.
Why Elisity Chose a CLI Instead of MCP
One of the most notable aspects of the launch is what Elisity deliberately avoided.
Instead of adopting the increasingly popular Model Context Protocol (MCP) architecture for AI integrations, the company built a traditional command-line interface.
According to Elisity, MCP servers require persistent, privileged connections that expand an AI agent’s access surface by default. That model has attracted increasing scrutiny from security researchers, particularly as enterprises evaluate how much authority autonomous AI systems should receive.
By contrast, the CLI operates only when invoked, runs under the user’s existing credentials, and requires explicit human approval before any action that changes system state. Delete operations require an additional confirmation layer.
The company’s design decision also follows growing concerns around AI infrastructure security. In July 2025, researchers disclosed CVE-2025-6514, a critical vulnerability affecting the widely used mcp-remote connector. The flaw, rated 9.6 out of 10 on the CVSS scale, allowed malicious MCP servers to execute operating system commands on connected machines, marking one of the first documented remote code execution attacks targeting an MCP client.
As enterprises expand AI-powered automation, minimizing privileged, always-on connections is becoming an increasingly important security consideration.
Identity-Centric AI for Zero Trust Networks
The new CLI is tightly integrated with Elisity IdentityGraph™, the company’s identity intelligence platform.
Rather than identifying endpoints solely through IP addresses, IdentityGraph correlates multiple identity signals to determine what a device actually is, providing richer context for AI-powered security decisions.
That identity-first approach aligns with modern Zero Trust strategies, where access decisions depend on verified identities, device posture, and contextual information instead of static network locations.
Human oversight remains central to the workflow. AI can recommend actions, generate reports, or automate routine analysis, but administrators retain final approval before any network policies are modified.
Customers See AI as an Assistant—Not an Autopilot
Elisity also shared adoption figures for Elisity Intelligence, its optional AI assistant.
According to the company, the number of customer organizations using the feature increased 77% between April and June 2026. The highest usage has centered on policy management, device identification, and dashboard-based operational tasks.
Importantly, Elisity says the AI does not operate autonomously. Administrators determine which features are enabled, and every recommendation requires human approval before implementation.
Current enterprise users include GSK, Main Line Health, and MultiCare Health System, the latter recently recognized with a CSO Award.
Customer Perspective: Continuous Validation Over Periodic Audits
Jason Elrod, Chief Information Security Officer at MultiCare Health System, said the CLI enables organizations to continuously verify that Zero Trust segmentation policies remain effective instead of relying on scheduled security reviews.
According to Elrod, automated validation helps detect policy drift within minutes, allowing organizations to confirm that segmentation controls continue blocking lateral movement before attackers can exploit configuration changes.
For industries such as healthcare—where operational downtime can directly impact patient care—that continuous assurance offers tangible resilience benefits beyond traditional compliance reporting.
Why It Matters
Enterprise cybersecurity is entering a new phase where AI is becoming less about autonomous decision-making and more about augmenting security teams with intelligent automation under strict governance.
While competitors increasingly market autonomous security agents, Elisity is betting that enterprises prefer transparency over automation for its own sake. By exposing an open CLI instead of a proprietary AI layer, the company allows organizations to integrate AI into existing workflows while maintaining identity-based controls and human oversight.
As AI adoption accelerates across Security Operations Centers (SOCs), the broader industry debate is shifting from whether AI should automate security tasks to how much authority those systems should receive. Elisity’s latest release positions the company firmly on the side of controlled, auditable AI—an approach that may resonate with highly regulated industries where accountability matters as much as automation.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
