According to a product walkthrough from The Hacker News, Bitdefender's GravityZone AI Visibility and Control (AIVC) is designed to answer a basic question: which AI software is running on protected Windows endpoints, who is using it, and how risky is it. The tool covers four exposure classes that typical software inventory misses: browser-based AI services, coding assistants inside development environments, locally running model runtimes and MCP servers, and AI tools configured to reach databases, repositories, or internal APIs.
AIVC discovery happens during the risk scan. The agent collects metadata about what AI software is present and how it is configured, but not what users do with it. The walkthrough stresses that AIVC does not inspect prompts, chat history, model weights, credentials, or full file content. It recognizes more than 50 AI agents and more than 70,000 skills, and assigns each discovered service to one of 17 functional categories such as LLMs, coding, productivity, and automation.
The console offers two views. AI Services is the full inventory of discovered tools with risk scores, categories, affected endpoints, and mitigation guidance. AI Findings is the prioritized list of what to look at first, with endpoint-specific evidence. The walkthrough notes that a score of 0 means no risky configuration was found, not that the organization has approved the tool, so administrators should review AI Services against their sanctioned list rather than relying only on findings. The sharpest example is an MCP server added through a few lines in a config file: no installer, no network traffic, but it appears in AI Services along with the assets it can reach.