At SOC Forum 2026, GK Solar introduced Solar AI Firewall, a new solution for filtering requests to external AI models, defending against prompt injections, and blocking potentially dangerous actions by AI agents.
The product is positioned for small, medium, and large enterprises that increasingly rely on AI assistants. While an AI agent may receive a legitimate task, this does not justify granting it unrestricted permissions. Solar AI Firewall is embedded directly into the data exchange between users and AI systems, applying multi-layer protection that operates in milliseconds and addresses 200 attack types.
Particular attention is given to autonomous AI agents that can independently execute actions inside corporate environments. The firewall continuously validates agent operations and blocks those that could lead to security incidents. This capability is especially relevant when AI tools move beyond recommendations and begin interacting directly with buttons and interfaces in enterprise systems.
The solution supports integrations with existing security infrastructure, including DLP, SIEM, and IDM platforms. It can be deployed either as a cloud service or in an on-premises configuration.
Solar has already implemented and tested the product within its own organization. Internal trials revealed the scale of shadow AI usage occurring outside established corporate governance. The vendor intends to expand the technology toward comprehensive, end-to-end management of AI-related risks.
