Based on an on-site inspection at RSAC 2026, this white paper outlines the necessity of redesigning security to adapt to the rapid growth of AI agents. It presents key trends and countermeasures for transitioning from traditional perimeter defense to zero trust and governance, including addressing Shadow AI and establishing identity and privilege management for AI agents. For the Japanese market, the report recommends a phased approach that integrates AI deployment environments with security controls.