Stellar Cyber's Auto-Triage AI Matches Human Analysts 99.7% Accurately

Post Views: 3

Stellar Cyber’s Agentic Auto-Triage system achieves 99.7% alignment with human analysts in evaluating security alerts according to an independent evaluation of 124 days of customer trials.

Independent Evaluation of 124 Days of Customer Trials

The study analyzed 138,475 real-world security alerts and produced identical conclusions to human analysts in 99.7% of cases. This research addressed a critical question for security teams considering autonomous SOC solutions: can artificial intelligence reliably make operational decisions?

Rising Cyber Threats and Generative AI-Enhanced Attacks

The findings come amid a surge in sophisticated cyber threats, with the World Economic Forum reporting a 48% year-over-year increase in ransomware attacks and a 1200% rise in phishing attempts since late 2022. These trends are attributed to generative AI-enhanced attack methodologies.

Agentic Auto-Triage Capabilities and Outcomes

The Agentic Auto-Triage capability automates the ingestion, correlation, analysis, and prioritization of suspicious activity.

Key Outcomes from the Trials

Key outcomes from the trials included a 19-minute productivity gain per analyst hour, equivalent to one day of work per week. This time savings enabled analysts to focus on higher-value tasks such as exposure management, threat anticipation, and proactive risk mitigation.

The system reduced noise by resolving 64% of false positives and escalating 15% of alerts as true positives for human review. The remaining alerts were categorized as informational, preventing unnecessary workload for security teams.

Quotes from Industry Leaders

Aimei Wei, CTO at Stellar Cyber, emphasized that security operations have reached a critical threshold where human analysts alone cannot manage alert volumes. “Our approach combines machine-speed analysis with human judgment,” Wei stated. “The AI handles large-scale alert processing while analysts maintain control, achieving near-perfect alignment and freeing teams to address more cases proactively.”

Christopher M. Steffen, VP of Research at EMA, highlighted the system’s ability to filter noise and focus on genuine threats. “Auto Triage’s AI-driven prioritization transforms SOC operations from alert processing to case management,” he noted. “This shift improves mean time to detect and respond while enabling analysts to adopt adversary-centric thinking.”

Chant Vartanian, CEO of M-Theory Group, described the results as transformative for managed security service providers. “Reclaiming a full day of productivity per analyst and eliminating 64% of false positives directly impacts service margins and client coverage,” Vartanian said. “This allows teams to focus on high-value threat investigations without expanding headcount.”

Availability and Future Presentation

The Auto-Triage feature is now available as part of the Stellar Cyber AI-native SecOps platform. The company will present the technology and study results at Black Hat USA 2026, showcasing its potential to redefine security operations.