On September 29, 2026, TrendAI, the enterprise business unit of Trend Micro Incorporated, launched a private preview of TrendAI Vision One Autonomous Vulnerability Discovery. The service identifies, validates, and verifies fixes for vulnerabilities, including zero-days, in customer source code. It pairs agentic AI with threat intelligence from the TrendAI Zero Day Initiative, the world's largest vendor-agnostic bug bounty program.
How the Technology Works
The service is powered by an agentic exploit-remediation engine under the code name AESIR and runs on a purpose-built, multi-model AI architecture that primarily uses Anthropic models. It integrates autonomous vulnerability discovery, exploit identification, and threat hunting with research from the TrendAI Zero Day Initiative, also known as ZDI. When evaluated against the CyberGym testing framework, the engine outperformed single-model and agentic system scores on the CyberGym leaderboard.
The workflow begins with a scoping interview to understand the customer's environment, followed by a custom scan defined and tuned by a TrendAI ZDI expert. Service teams monitor and control every scan rather than leaving customers with an unsupervised black box. After scanning source code, the system produces a comprehensive vulnerability report that includes severity ratings, estimated CVSS scores, and detailed write-ups for every finding.
Key Differentiators
TrendAI highlights three differentiators for the service. Internal benchmarking shows that its AI models identify more true-positive vulnerabilities and fewer false positives than leading vulnerability scanning and static analysis tools. Findings are also validated by ZDI experts against more than 20 years of research from the program and the team behind the Pwn2Own hacking competition.
The service is deployed and operated on Amazon Bedrock from Amazon Web Services, allowing customers to retain their existing Git repositories and code registries as the system of record. This means source code remains within the customer's AWS environment throughout the entire workflow, from initial scan through final patch verification. The approach supports security requirements, operational continuity, and existing development practices.
Scan and Verification Process
After a scan runs against customer source code, TrendAI cross-checks every discovery against ZDI vulnerability research. Customers can then build and deploy a patch, after which the scan is run again to verify the fix. The result is a final verified patch status that is again checked against ZDI research before delivery.
Executive Perspective
Rachel Jin, Chief Platform and Business Officer and Head of TrendAI, said, "AI is changing both sides of vulnerability discovery." She added that security teams should not have to choose between the speed of AI and the judgment of the world's best vulnerability researchers. The new service puts AI built for accuracy together with 20 years of ZDI research and runs it where customers' code already lives.
Availability
Autonomous Vulnerability Discovery is available now in private preview to select enterprise customers running on AWS. The controlled preview allows TrendAI experts to tune each deployment to customer environments, reflecting the service's native integration with Amazon Bedrock. Pricing is customized based on scan scope and deployment model, and customers can request a scoping interview through their TrendAI account team.
The announcement reflects a broader shift toward AI-assisted security operations that combine automation with expert validation. By embedding the service inside customer AWS environments and pairing it with ZDI research, TrendAI aims to help organizations reduce exploitable vulnerabilities faster. Private preview participants will help shape the product ahead of wider availability.