Generative AI tools are now part of daily work — and a new data-egress channel. Employees paste source code, customer records and strategy documents into public chatbots to get help, often with no idea the data may be retained.
Why this is different
Unlike email or USB, the GenAI channel is a browser text box. Legacy DLP built for files and attachments never anticipated it, so most sensitive pastes go completely unseen.
Detecting sensitive prompts
From the Pro edition, Verikor inspects web and HTTP(S) traffic — including text entering browser-based and AI tools — and applies allow, warn or block based on classification and content. Staff keep the productivity; the crown jewels stay in-house.
Enablement, not prohibition
Blanket bans push AI use into the shadows. A warn-and-guide approach lets people use AI safely while the genuinely sensitive content is caught in real time.