Skip to content
KavachLab

Resources

Enterprise AI governance, explained

Practical guides and plain definitions for the teams responsible for AI risk — written for security, privacy, and compliance leaders rather than for a search engine.

Glossary

The terms that come up in every enterprise AI governance conversation.

Shadow AI
Unsanctioned use of AI tools, models, or extensions by employees, outside the visibility and approval of the security team.
AI-aware DLP
Data loss prevention that inspects and redacts sensitive content inside AI prompts and uploads before the data reaches a model, rather than inspecting network traffic after the fact.
Inline enforcement
Applying a policy decision synchronously at the moment of interaction — before data is sent — instead of detecting a violation retrospectively in logs.
Customer-hosted data plane
An architecture where all data processing runs inside the customer’s own tenant, so the vendor never takes custody of prompts, files, or monitoring data.
Prompt injection
An attack where crafted text causes a model to ignore its instructions, leak data, or take unintended actions on behalf of an attacker.
Explainable risk score
A risk rating that exposes the weighted factors behind it, so an analyst can see why an interaction was scored as it was rather than trusting a black box.

Ready to see every AI interaction?

Talk to our team for a guided demo and a scoped, monitor-only discovery pilot.

Request a Demo

Fields marked with an asterisk are required.

Which AI tools are in use?

By submitting, you agree to our Privacy Policy.