Global IT corporation FPT has launched AI-powered enterprise cybersecurity solutions to bring security remediation up to the pace of AI-driven attacks. Combining expert-led oversight with AI models, the solutions help identify and prioritize real exploitable threats across critical assets, reduce risk immediately through targeted protection, and build continuous defense capabilities over time.

With an 89% year‑over‑year surge in AI‑powered attack attempts, frontier models are rapidly lowering the barrier for attackers to discover and exploit vulnerabilities at scale that were previously missed by traditional tools. This shift not only increases the volume of new vulnerabilities across all organizations but also compresses the time to weaponization, with exploit windows shrinking from years to just days. Enterprises are struggling to remediate fast enough, resulting in a growing backlog of security weaknesses and unpatched critical exposures. This requires organizations to remediate vulnerabilities at scale in the short term while building AI security readiness to keep pace with increasingly compressed threat cycles.


FPT showcased AI-powered cybersecurity offerings, ranging from intelligent security monitoring and real-time threat detection to penetration testing, risk assessment, and 24/7 managed security services.

Combining AI-powered cybersecurity solutions and existing security tools with expert-led oversight, FPT introduces its risk-based remediation approach, helping businesses rapidly fix critical weaknesses and harden against exposures.  Rather than surfacing every possible finding, the approach identifies and prioritizes exploitability-validated risks tied to critical assets, enabling organizations to fix what matters most. Systems are first classified based on exposure, business criticality, and data sensitivity to define scope. Under expert-led oversight, AI models work alongside existing security tools to help identify exploitable vulnerabilities on critical systems, with findings validated against real-world conditions to eliminate false positives.

In the remediation stage, FPT combines immediate protection measures with permanent fixes. In a three-to-six month phase, FPT deploys large-scale, AI-augmented resources to systematically identify and remediate vulnerabilities across multi-layers (infrastructure, applications, and cloud environments), which quickly closes the gap between attack speed and defensive capacity. Virtual patching blocks active attacks and reduces exploitable risk instantly, while AI-assisted remediation accelerates vulnerability resolution and clears backlogs at scale. For enterprises running custom-built applications, FPT extends remediation to the source code level using Flezi Metis, its proprietary AI code intelligence solution, enabling engineers to understand complex codebases in hours, pinpoint where fixes are needed, and resolve vulnerabilities at their root.

As systems evolve and new threats emerge, FPT enables continuous AI-powered detection, remediation, and response services, helping enterprises maintain resilience in the long term. FPT delivers these through two dedicated platforms designed for continuous operation.

  • FPT Shugo Omizu, an autonomous AI agent-based penetration testing platform, replaces periodic assessments with continuous, automated vulnerability discovery. It cuts onboarding time from up to 10 days to 2 hours, reduces execution time from weeks to under 50 hours, and completes remediation validation in under 60 minutes.
  • FPT Shugo Kyro, the AI-powered engine behind FPT's Managed Extended Detection and Response (MXDR) service, delivers integrated detection and response across environments. It acknowledges all alerts within 10 minutes, responds to high-severity incidents within 13 minutes, and automatically resolves up to 70% of alerts under expert supervision.

Through FPT's AI-powered cybersecurity solutions, enterprises have been able to significantly reduce their exposure to attacks and respond to threats faster. These outcomes have been demonstrated across multiple enterprise scenarios, from managed security assessment for Japanese enterprises, to improving visibility and automated patching in multi-cloud environments, and helping security teams reduce alert noise and accelerate threat response.

“AI can already discover and exploit vulnerabilities faster than most enterprise defenses can respond. The time to identify and harden critical exposures is now, before powerful offensive AI becomes broadly available. FPT responds to this with an end-to-end approach, where expert-led oversight ensures the right priorities are set, and AI-powered solutions act on them with the speed and precision, enabling organizations to build more resilient platforms,” said Pham Tung Duong, FPT Software Vice President and Director of CyberSecurity Assurance Service, FPT Corporation.


FPT leverages an AI-augmented workforce of over 30,000 engineers, including more than 300 certified cybersecurity specialists, to ensure remediation at scale.

Underpinning these capabilities is FPT's One Team engagement approach, bringing together cybersecurity specialists and technical experts across disciplines. With teams ready to mobilize within days across all workloads, this cross-functional composition enables deeper system understanding for faster diagnosis, and provides full-spectrum coverage across every vulnerability type and environment.

FPT has established strategic partnerships with leading cybersecurity organizations, including Amazon, Microsoft, Cybereason, and CrowdStrike, to enhance enterprises' protection against cyber threats. The company is also a member of CREST and the Microsoft Intelligent Security Association, underscoring its commitment to maintaining high standards of cybersecurity expertise. With extensive experience across cybersecurity engagements, the company delivers end-to-end cybersecurity capabilities across offensive security, Security Operations Center, MXDR, and AI security consulting. In a recent one-month AI-assisted program, FPT identified critical vulnerabilities across widely used platforms, including Llama.cpp, Kysely, Kestra, and leading WordPress plugins, and demonstrated advanced exploitation capabilities with a successful exploit of NVIDIA Triton Inference Server at Pwn2Own AI 2025.