Astelia Researcher Says AI Agents Multiply Cybersecurity Output Tenfold

Astelia founding research engineer Yuval Barak said AI agents have raised each researcher's output to roughly ten times its previous level. He said the agents take on laborious tasks such as enumerating configurations, reading documentation, reproducing behavior across variations and maintaining notes, while people still select investigations and verify the machine's work. Astelia's platform models network topology, segmentation, controls and asset reachability to prioritize vulnerabilities it considers exploitable. The company's first-party product description says its agents carry findings from triage through remediation and validation, support more than 100 MCP integrations, log actions for audit and retain human approval checkpoints.