Cytex Unveils AICenturion: The Industry's First LLM Firewall with Integrated DLP Securing the Use of AI for Public and Private Sector Entities

WASHINGTON and AUSTIN, Texas, March 18, 2025 /PRNewswire/ -- Cytex, Inc. a trailblazer in AI-driven cybersecurity, announced the launch of AICenturion, a groundbreaking LLM Firewall with Data Loss Prevention (DLP) capabilities.  GenAI's risks intensify as LLMs prevent enterprises from directly controlling their processes and data handling.  AICenturion provides the trust, risk and security management that enterprises need by enabling AI model discovery, data loss prevention, threat mitigation, and compliance framework policy enforcement.

"With the explosive adoption of LLMs, we're seeing a new frontier of risk emerge: the exfiltration of sensitive data uploaded into LLMs, data leaks, prompt injections, and model theft," said Andrew Surwilo, CEO of Cytex. "AICenturion isn't just a tool; it's a strategic enabler. It allows businesses to leverage AI while maintaining defense against threats that traditional security solutions simply can't see. Our mission is to make AI a trusted partner, not a liability." 

AICenturion provides risk mitigation of LLM vulnerabilities:

• Detection of LLM and AI models across the enterprise
• Policy enforcement for different roles and adherence to acceptable use
• Monitor usage against stated objectives and adjust usage parameters
• Coverage of the OWSAP top LLM vulnerabilities:
  
  
  • Sensitive Data Disclosure (LLM01): Identify and block PII, PHI, and CUI from leaking through LLM prompts or outputs.
    
    
  • Prompt Injection (LLM02): Review prompts for data exfiltration and malicious injection attacks.
    
    
  • Insufficient Access Controls (LLM03): RBAC policy enforcement ensures that users are only allowed to view data that is consistent with their role.
    
    
  • Insecure Output Handling (LLM04): Monitor and filter LLM outputs for sensitive and malicious content. Automated data redaction ensures LLM generated risky content, and malicious content is filtered.
    
    
  • Overreliance on LLM Outputs (LLM06): Limits LLM use to authorized scenarios and users according to their respective roles, reducing the chance of overreliance on unverified outputs.
    
    
  • Model Theft (LLM07): A fortified firewall along with Cloud Security Posture Management and usage pattern analysis thwart exfiltration attempts targeting LLM intellectual property.
    
    
  • Excessive Agency (LLM08): Role-based policies ensure LLMs only act within approved boundaries, curbing unauthorized decisions.
    
    
  • Supply Chain Vulnerabilities (LLM09): Cytex platform's SBOM and SCAT analysis integrated within the development processes ensure vulnerabilities within the traditional software supply chain and model plugins are identified.
    
    
  • Denial of Service (LLM10): Proactive mitigation leveraging application security techniques with Cytex detects and thwarts overload attempts, limiting the attack surface.

Safeguarding Internal AI Tools
AICenturion goes beyond external threats by protecting organizations from risks posed by internal agentic AI tools. Through its policy enforcement and contextual awareness, AICenturion restricts access to sensitive data based on user roles preventing unauthorized internal users, from rank-and-file employees to executives, from inadvertently exposing confidential information via AI interactions. Whether it's blocking Social Security numbers from being processed by an AI model or ensuring compliance with organizational policies or data privacy frameworks, AICenturion keeps internal AI tools in check.

AICenturion along with the Cytex Unified Resilience platform's runtime protections make it an ideal fit for organizations using sanctioned LLMs in production environments.

"AICenturion redefines how we secure AI at the intersection of innovation and risk," said Taimur Aslam, CTO of Cytex. "Our real-time monitoring and DLP capabilities act like a digital immune system for LLMs – constantly scanning, filtering, and adapting to threats, whether from external attackers or internal tools like Copilot. With Leo's intelligence layered on top, we're not just reacting to dangers; we're anticipating them, giving our clients a decisive edge in the AI era."

The launch comes at a pivotal moment. As LLMs power everything from customer service to strategic decision-making, the stakes for security have never been higher. AICenturion's ability to enforce compliance, protect intellectual property, prevent overreliance on unverified outputs, and secure internal AI deployments positions it as a must-have for CISOs, compliance officers, and IT leaders navigating this uncharted territory.

Cytex invites organizations to experience AICenturion firsthand. For more information, visit http://Cytex.io/iacenturion or contact Cytex to schedule a demo at  [email protected].

About Cytex
Cytex is a leading innovator in AI-powered cybersecurity, delivering solutions that protect enterprises from emerging threats in an AI-driven world. With a focus on real-time protection & monitoring, data security, and compliance, Cytex empowers organizations to embrace transformative technologies with confidence.

SOURCE Cytex, Inc.