---
title: "Data Loss Prevention (DLP)"
description: "DLP inspects outbound data flows to prevent sensitive content from leaving. Why traditional DLP doesn't catch AI prompts, and what fills the gap."
url: "https://sonomos.ai/glossary/data-loss-prevention/"
source: "https://sonomos.ai/glossary/data-loss-prevention/"
---

# Data Loss Prevention (DLP)

A category of security tooling that inspects outbound data flows to prevent sensitive content from leaving an organisation's control.

## In depth

Data Loss Prevention (DLP) is the umbrella term for tools that watch outbound traffic — email, file shares, cloud uploads, USB transfers — and block, alert, or transform when content matches a sensitive pattern. Traditional DLP runs as a network gateway, an endpoint agent, or a sanctioned-app integration. It is mature for structured data flows. Its weakness in the AI era is that the dominant exfiltration channel is now a free-form chat box that often does not pass through any of those choke points, and the inspection point is server-side — meaning the data has already left the device.

## Examples

- A network DLP gateway blocking an email attachment that contains 50+ Social Security Numbers.
- An endpoint agent quarantining a USB transfer of customer database exports.
- A SaaS DLP integration redacting credit card numbers in a Salesforce export.

## How Locke handles data loss prevention (dlp)

Locke is DLP designed for the AI prompt path. It runs on the endpoint (on your device), inspects the prompt before transmission, and acts at field level (mask or block). It complements rather than replaces traditional DLP: existing DLP for email and file shares, Locke for AI prompts.

[See the product](https://sonomos.ai/locke/)

## Related terms

- Privacy Layer for AI

  A control point between a user and an AI service that detects, transforms, or blocks sensitive data before it reaches the model.
- Prompt Leakage

  The unintended disclosure of sensitive information through the content of a prompt sent to an AI model.
- Shadow AI

  Employees using AI tools at work without organisational sanction — the AI-era version of shadow IT.

## Reading on this topic

- [The Best AI Privacy Tools for ChatGPT, Claude, and Gemini (2026)](https://sonomos.ai/blog/best-ai-privacy-tools-chatgpt-claude-gemini-2026/)
- [AI Data Leakage: 7 Ways Sensitive Information Escapes to LLMs (and How to Stop It)](https://sonomos.ai/blog/ai-data-leakage-7-ways-sensitive-info-escapes-to-llms/)