Cyware Deploys AI Agents to Automate Cyber Defense

NEW YORK, NY – March 18, 2026 – Cybersecurity firm Cyware today announced a significant expansion of its artificial intelligence strategy with the launch of its Agentic Fabric, a new platform designed to deploy autonomous AI "agents" to assist security teams. The move signals a broader industry shift from AI as an analytical tool to AI as an active, goal-oriented participant in cyber defense.

Building on a foundation introduced in late 2025, the Agentic Fabric aims to create threat intelligence-driven workflows that automate complex tasks across security investigation, detection, and response. These are not simple scripts or playbooks; they are goal-driven AI agents capable of reasoning, planning, and executing multi-step cybersecurity operations with a degree of autonomy previously unseen in mainstream security products.

"Cyware is aggressively advancing threat-centric Agentic AI for cyber defense to meet pent-up customer and industry demand," said Sachin Jade, Chief Product Officer at Cyware, in a statement. "Agentic Fabric introduces an ecosystem of specialized agents that work alongside analysts, applying threat intelligence and correlated data and executing objectives across the entire security lifecycle. This approach empowers defenders to innovate at the pace of the adversaries they face."

The Rise of the Digital Teammate

The launch taps into the burgeoning field of "agentic AI," widely considered the next evolutionary step after the generative AI boom. While generative AI excels at creating content in response to prompts, agentic AI takes it further by using AI-generated plans to autonomously interact with tools and environments to achieve a specific goal. In cybersecurity, this translates to AI systems that can function as digital teammates for overburdened security analysts.

Cyware's platform is structured around an "Analyst Agent Hub," a central console where human analysts can coordinate and deploy these AI agents. The system features a growing "Agent Catalogue" of specialized digital workers, each designed for a specific function:
* The Attack Flow Agent reconstructs adversary timelines and maps their activities to industry-standard frameworks like MITRE ATT&CK.
* The Contextual Intelligence Agent ingests raw threat data and summarizes it into concise, human-readable intelligence.
* A SOC Analysis Agent assists with the triage and investigation of security alerts, providing context on why an event might be a threat.
* The Detection Engineering Agent helps analysts create new threat detection logic, capable of writing YARA or Sigma rules on the fly to counter emerging threats.

This model represents a new paradigm for human-AI collaboration in the Security Operations Center (SOC). By offloading time-consuming data collection, correlation, and initial analysis, the agents are designed to free up human experts to focus on higher-level strategic thinking, complex threat hunting, and final decision-making. The goal is to augment, not replace, human capability, turning a flood of data into actionable insights at machine speed.

A Crowded Battlefield for AI Supremacy

Cyware's announcement does not happen in a vacuum. The cybersecurity industry is in the midst of an intense race to integrate more advanced AI, with "agentic" capabilities becoming the new competitive frontier. Gartner has identified agentic AI as a top strategic technology trend, noting a staggering 1,445% increase in enterprise inquiries about multi-agent systems between early 2024 and mid-2025. The cybersecurity-specific agentic AI market, valued at over $22 billion in 2024, is projected to soar past $300 billion by 2033.

Established security giants and nimble startups alike are vying for dominance. CrowdStrike has promoted its own AI agent, "Charlotte AI," as a conversational security analyst. Palo Alto Networks is focusing on both deploying and securing agentic systems. Meanwhile, new players are emerging with massive financial backing; the startup Kai, for instance, launched in March 2026 with $125 million in funding for its own agentic AI platform aimed at unifying security workflows.

This surge is driven by a market realization that while generative AI was a powerful new interface, true business impact comes from autonomous execution. For security teams facing a severe talent shortage and an overwhelming volume of alerts, the promise of autonomous agents that can independently manage and resolve incidents is incredibly compelling.

An Arms Race in the Digital Trenches

The push toward agentic defense is also a direct response to a disturbing trend: adversaries are weaponizing AI. Malicious actors are already using AI to accelerate reconnaissance, craft more convincing phishing emails, generate polymorphic malware, and discover vulnerabilities at an unprecedented rate. This creates an asymmetry where attackers can operate at machine speed while defenders are still constrained by human-led processes.

Agentic AI is being positioned as the essential countermeasure to level the playing field. The technology promises to shift security postures from reactive to proactive. Instead of just responding to alerts, these systems can continuously hunt for threats, analyze network behavior, and operationalize threat intelligence in real time. Cyware's emphasis on "threat intelligence-driven workflows" is key here, suggesting a system where agents can autonomously ingest new intelligence about an adversary's tactics and immediately adapt the organization's defenses, such as by generating and deploying new detection rules without human intervention.

This represents a fundamental change in the speed and scale of cyber defense, moving from a model of human-led incident response to one of continuous, autonomous validation and remediation. The ultimate vision is a self-defending network that can outmaneuver AI-powered attacks in real time.

Navigating the Risks of Autonomous Defense

Despite the immense promise, the rapid deployment of autonomous agents in critical security functions introduces a new and complex risk landscape. The very power that makes these agents effective—their ability to act independently—also makes them potentially dangerous if not properly controlled. The industry is grappling with significant challenges, including the "black box" problem, where the complex reasoning of an AI agent can be difficult for humans to understand, audit, or trust.

An incorrect decision by an autonomous agent could have severe consequences, from generating a flood of false positives that waste analyst time to, in a worst-case scenario, incorrectly shutting down a critical production server during a misidentified incident. Security for the agents themselves is another major concern. These systems create a new attack surface, with risks like "prompt injection" or "automation hijacking" that could allow an adversary to manipulate an organization's own defensive agents and turn them against it. Forrester has gone so far as to predict that 2026 will see the first public data breach caused by a compromised agentic AI system.

This new reality necessitates a "human-in-the-loop" approach, where human experts retain ultimate oversight and control, especially for high-impact actions. Building robust governance, clear accountability frameworks, and reliable kill-switches will be just as important as developing the agents' capabilities. The race is now on not only to deploy these powerful new digital teammates but also to build the guardrails and security standards needed to ensure they remain trusted allies in the fight against cyber threats.