AI discovery gap calculator
You cannot govern what you cannot see. Size the share of your AI estate that security has never inventoried.
Every governance program starts with a list of the AI in use — and in most enterprises that list is a fraction of reality. Employees adopt tools on personal accounts, teams wire up agents with tool and MCP access, and none of it files a ticket. This calculator sizes the gap between the AI estate you have inventoried and the one the network can actually see.
01: What security currently knows about
02: The agentic layer
Agents act on their own, through tool calls and MCP connections—each one is an access path.
Inventoried vs dark
Each cell is 1% of the AI actually in use. Blue is on the sanctioned inventory. Hollow red is in use, ungoverned, and invisible to every control that depends on knowing it exists.
More Info
Discovery is the step governance programs skip because it is hard: you cannot ask people to register tools they do not think of as tools, and agents never ask at all. Network-based discovery inverts the problem—instead of relying on declarations, it observes AI traffic where it actually flows, classifies the intent behind it, and traces agent tool calls and MCP connections in real time, with no application changes.
Frequently asked questions
From AI factory math to AI factory control
Closing the gap takes network-based AI discovery, intent classification, and shadow AI detection in the traffic path. Discover, test, and govern—in that order, because the order only works one way.