DLP answers a question every enterprise eventually faces: how do you stop your own data from leaving in ways it shouldn't? The technology inspects content — emails, file uploads, clipboard contents — for sensitive patterns and enforces a policy when it finds them.

Two recent grants make the mechanism concrete and show where it's heading. Zscaler's US12603868B2, "Endpoint data loss prevention" (issued April 14, 2026; CPC H04L 63/04), describes DLP enforcement at the endpoint itself — on the device where data is being used. Netskope's US12316647B1, "Video data loss prevention (vDLP)" (issued May 27, 2025), extends the same idea to a surface most DLP never covered: video and screen content. Read them at US12603868B2 and US12316647B1.

“Systems and methods include receiving Data Loss Prevention (DLP) configurations for one or more devices, wherein the DLP configurations define how exfiltration of sensitive data is protected for the one or more devices; monitoring traffic of the one or more devices; and scanning the traffic of the o…”— U.S. Patent No. 12,603,868 source

The way this actually works is inspect-classify-enforce. The system observes data at a control point (an endpoint agent, a network proxy, a cloud API), classifies whether it contains something sensitive — a credit-card number, a source-code file, regulated personal data — and then applies the policy: allow, log, warn, or block. The Netskope vDLP grant is notable because it pushes that inspection into video frames, acknowledging that data leaks through screenshots and screen-shares, not just file transfers.

Here's the business angle this desk cares about. The progression in these two grants — from files to endpoints to video — is the product story of the whole DLP market: the set of surfaces to inspect keeps expanding, because the ways people move data keep expanding. Every new work pattern (remote screen-shares, AI copy-paste, mobile) opens a new leak path, and each one is a reason to expand the DLP footprint. That expanding surface is what turns DLP from a one-time purchase into recurring, growing demand.

The honest caveat: DLP is famously noisy, and a patent on an inspection method is not a guarantee of low false positives. The grants describe how detection and enforcement work, not how accurate they are in production. Buyers should read the mechanism as a capability claim, not a quality claim.

The grounded read: DLP is content inspection plus policy enforcement at a control point. When a vendor adds a new "DLP for X," what's really changing is the control point — a new surface to inspect. These two grants name two of those surfaces explicitly, and the direction they point is the same one the market is buying: cover every place data can move.