WebKeeper AI DLP
Data Loss Prevention for Confidential, Personal, and Sensitive Information in Generative AI
WebKeeper AI DLP is a unique network security solution that brings together SOMANSA’s 30 years of expertise,
purpose-built to prevent data leakage in generative AI environments.
It detects and controls personal and confidential information exposed through generative AI,
using LLM-based semantic analysis to prevent both accidental and unauthorized transmission while logging all access activity.
A standard solution for controlling confidential data leakage through generative AI, optimized for the N2SF national network security framework and
first deployed by a major Korean financial institution.
Market-Leading AI DLP for Preventing Confidential, Personal, and Sensitive Data Leakage Through Generative AI
From Invisible Shadow AI Access to Confidential Data Leakage
Complete Control of Generative AI with WebKeeper AI DLP
Built on SOMANSA’s 30 years of dedicated DLP expertise,
WebKeeper AI DLP protects your organization from personal data leakage in generative AI environments.
Experience powerful AI DLP that protects the entire lifecycle—from detection and blocking to control and post-incident audit.
30 Years
Network Security Expertise
60+
AI Services Supported
80%
Adoption Rate among Korean Financial Institutions
CORE STRENGTHS
Record Every Interaction with Generative AI
and Uncover Hidden Data Exfiltration Intent.
Use Generative AI Freely. Keep Corporate Data Secure.
From detection and blocking to audit, WebKeeper AI DLP controls every potential data leakage scenario involving generative AI.
01 Complete Visibility
Store All Inputs and Responses
Reconstruct Conversations in Chat Format
Beyond simple blocking, prompt inputs and AI responses are reconstructed in the actual chat format.
Comprehensive logging clearly shows who entered what and when, helping identify signs of data leakage.
Original transmitted files are preserved to support efficient post-incident audits and tracking.
02 Precision Detection & Control
Block Images, Drawings, Blueprints, and Context-Based Content
Real-Time Blocking of File/Text Transmission
Combining Korea’s largest personal data checksum library with LLM-based contextual analysis, WebKeeper AI DLP precisely detects and blocks personal and confidential information contained in files and prompts.
It analyzes the meaning and intent of unstructured data—including sentences, paragraphs, documents, and images—to provide sophisticated data leakage controls beyond simple keyword detection.
03 Personal Data Pattern Detection
30 Years of Network Security Expertise
Precision Detection of Personal Data Patterns in Generative AI Environments
For 30 years, SOMANSA has built security environments for public-sector, financial, and enterprise organizations, accumulating extensive expertise and technology.
Based on Korea’s largest personal data checksum library, it blocks personal data leakage through generative AI.
We extend proven security technologies to generative AI environments to provide more precise personal data leakage control.
From Input to Response
WebKeeper AI DLP Controls the Entire Flow
WebKeeper AI DLP Architecture
KEY FEATURES
What Are Your Employees
Entering into Generative AI?
WebKeeper AI DLP controls the use of generative AI across your organization and logs all prompts, responses, and transferred files.
If personal or confidential information is detected, external transmission is blocked to ensure a secure AI work environment.
ACCESS CONTROL
Policy-Based Access Control
Granular policies by department, service, and time of day
provide precise control over generative AI usage.
Department Policies
Access Permissions by Organization
Service Policies
Allow/Block by Service
Time-Based Policies
Access Control by Time of Day
CONTENT-BASED BLOCKING
Content-Based Blocking
Analyzes and blocks suspected personal or confidential data leakage in real time.
Regex
Detect Personal Identifiers, Bank Account Numbers, and Card Numbers
Keyword
Detect Restricted Terms and Confidential Keywords
Embedding Analysis
Document, Sentence, and Image Similarity Analysis
LLM Analysis
Context-Based Risk Assessment and Document Classification
AUDIT TRAIL
Unified Conversation History
Reconstruct prompts, responses, and transmission history in a chat format for an intuitive view of employees’ AI usage.
- Records users, dates, purposes, and input/output content
- Stores all user prompts and AI responses
- Supports post-event audits based on transmitted file logs
AI DOCUMENT CLASSIFICATION
AI-Powered Document Category Classification
Analyzes document content and automatically classifies it into 20 major categories and around 100 subcategories,
then assigns one of three security levels based on sensitivity and risk.
| Category / Subcategory | Type | Level |
|---|---|---|
| Management / Strategy | ||
| Business Plan | Document | High |
| Executive Report | Document | High |
| Research / R&D | ||
| Research Report | Document | High |
| Design Specification | Drawing | Medium |
and more
COVERAGE & COMPLIANCE
Control a Wide Range of Generative AI Services
and Meet Key Security Standards, Including N2SF
Control 60+ generative AI services and
meet N2SF-based security requirements with a single solution.
SUPPORTED MODELS
Coverage for 60+ Generative AI Services
Supports more than 60 generative AI services, including ChatGPT, Claude, and Gemini.
Control both major and newly emerging AI services, with coverage continuously expanding.
ChatGPT
Gemini
Copilot
Cursor
Grok
Perplexity
Deepseek
Claude
Wrtn
Clova X
ChatPDF
BingChat AI
Notion AI
Github Copilot
Canva AI
COMPLIANCE
Meet N2SF-Based Generative AI Security Requirements
A solution that meets the National Network Security Framework (N2SF) guidelines established by Korean government agencies.
In line with the N2SF policy rollout, SOMANSA has proactively deployed the solution across government and public-sector organizations, establishing stable operating environments.
| Application Area | Key Security Requirements | Technical Safeguards |
|---|---|---|
| Business Endpoint Environment | Endpoint Network Security | AI DLP, Endpoint DLP, VDI, SWG |
| AI-Integrated System Environment | User and Device Authentication for Generative AI Services | AI DLP, Account Management System |
| AI-Integrated System Environment | Data Protection When Using Generative AI | AI DLP (Real-time blocking of personal data transmission, audit logs retaining original transmitted files) |
| Internet Environment | Data Management for Generative AI Usage | Audit Logging and Monitoring for Generative AI Services |
REFERENCES
Key References
For 30 years, SOMANSA has led the fields of internet access control and customer data leakage detection, backed by accumulated security expertise and R&D capabilities.
Chosen by more than 50 customers in Korea and abroad, including the National Tax Service, Board of Audit and Inspection, LG U+, KakaoBank, and major financial institutions, SOMANSA delivers security architectures optimized for each customer’s environment based on extensive deployment experience.
GET STARTED
For the Safest Generative AI Adoption,
Start with SOMANSA.
Frequently Asked Questions
We’ve compiled the questions most frequently asked before deploying WebKeeper AI DLP.
First question.
Answer to the first question.
Second question.
Answer to the second question.
Third question.
Answer to the third question.
