Skip to content
NetrikSolutions
Discover · Detect · Protect · Govern

Know exactly where your personal data lives.

Argon discovers, classifies and helps protect PII across your databases and file stores — combining a 2,000+ pattern library with a locally hosted LLM that reads context, not just format. Read-only discovery, with local detection that keeps scan data on your infrastructure.

Argon / Privacy workspace
Argon dashboard showing four sample data sources, 240 sample findings, severity distribution and governance readiness
Actual Argon interface · Illustrative workspace with sample data

Read-only

The scanner never needs write access to a source.

On your infrastructure

On-premises, private cloud or VPC. Air-gapped supported.

Local LLM

Context analysis runs inside your perimeter, not a vendor API.

2,000+

Built-in PII detection patterns

3

Detection modes: Regex, LLM, Hybrid

Global

Regulation-aware governance

Local-first

Detection on your infrastructure

The Argon platform

Privacy, with the full picture.

From connecting a source to preparing the evidence — see the actual workspace your team uses to discover, understand and govern personal data.

01 / 05The full picture

One view of your privacy programme.

Configured sources, scan activity, sensitive-data findings and governance readiness. A working overview that helps your team decide what needs attention next.

  • Connected sources and discovery coverage
  • Finding severity and identifier categories
  • Quick access to database, file, folder and SSH scans
Argon / Workspace
Argon dashboard with four sample sources, 240 findings and a severity breakdown
Actual Argon interface · Illustrative workspace with sample data
01Discover & classify

Bring hidden data into view.

Patterns identify candidates. Context analysis can help evaluate them. Your team gets source-level detail to investigate the findings.

02Govern with context

A programme that speaks your law.

Keep the same implementation phases while the controls, terminology and options change with the regulation you choose.

03Prepare the evidence

Give every activity a clear owner.

Connect processing purposes, departments, systems and retention to the sensitive data your organisation handles.

From connection to evidence — the complete workflow 8 capabilities
01 · Overview

One view of every configured source

Assets, scan activity, PII counts, compliance posture and licence state on a single operating screen.

02 · Connect

Guided source onboarding

Add enterprise RDBMS, warehouses, NoSQL and document sources, organised by data platform.

03 · Scope

Set the scope before you run

Launch focused scans with the target count, scan mode and pattern coverage visible up front.

04 · Detection

Choose Regex, LLM or Hybrid

Pick the detection mode and the pattern sets in the same workbench as the source scope. Mode availability depends on the runtime and source.

05 · Results

Triage what came back

Scan history, completion status, finding counts, severity and source context in one explorer.

06 · Breakdown

Impact by database and table

Affected databases, tables, confidence and identifier categories, ready for remediation planning.

07 · Report

Evidence with masked samples

Grouped findings showing masked values, severity, confidence, affected columns and rows impacted.

08 · Analytics

Roll-up for the people who ask

Severity breakdowns, PII distribution and database coverage across completed scans.

Discovery, classification and governance. On your infrastructure.

The challenge

Data silos are where personal data goes missing.

Sensitive data spreads across legacy systems and modern cloud databases faster than any team can track it. A scanner limited to one environment — or one that requires moving data to work — leaves exactly the gaps that matter. Effective discovery has to be database-agnostic and give one view across the whole estate.

Where sensitive data ends up

  • Legacy relational systems nobody has audited in years
  • Modern cloud databases and warehouses added team by team
  • File servers, NAS and network shares holding KYC and contracts
  • Copies made for testing, reporting or analysis and never removed

What that costs you

  • Personal data in locations nobody has inventoried
  • No evidence trail when a regulator or auditor asks
  • Excess data retained well past its purpose
  • No way to prioritise which exposure to fix first
Coverage

Database-agnostic, by design.

A single pane of glass over structured databases and unstructured file systems — so risk can be assessed across the whole organisation without migrating data or running a different tool per environment.

Relational databases

Deep interrogation of schemas, tables and records.

  • PostgreSQL
  • MySQL
  • Microsoft SQL Server
  • Oracle

NoSQL and cloud databases

Document and managed database services.

  • MongoDB
  • Managed cloud databases
  • Document stores

File repositories

Unstructured stores where KYC packs and contracts accumulate.

  • File servers
  • NAS
  • Network shares

Object storage

Bucket and blob storage across providers.

  • Object storage buckets
  • Archive tiers
Connector library

Across your data estate.

All 55 connector types from Argon: databases, warehouses, messaging, storage and cloud drives.

Connector availability, drivers, permissions and authentication depend on your deployment. Vendor marks identify the connected products; SAP and IBM products use their vendor marks.

Detection engine

Patterns find it. Context confirms it.

Scanning on regular expressions alone produces either a flood of false positives or quiet misses on obfuscated data. Argon pairs a large pattern library with an NLP layer that evaluates what surrounds a value before calling it sensitive.

Mode availability depends on your source and runtime. Server database scans use Regex; supported file, folder and SSH workflows can use AI analysis. Desktop database AI uses an eligible configured provider.

Regex

Pattern-based detection of known formats

Over 2,000 built-in patterns covering global financial identifiers, healthcare records and regional PII. Fast, broad first-pass coverage across a whole source.

LLM

Context-based identification

A locally hosted language model reads the surrounding context to judge whether a value is genuinely sensitive — the verification tier that pattern matching cannot provide.

Hybrid

Pattern and context together

Patterns find candidates at speed, the model confirms them in context. This is the mode most teams run for a balance of coverage and precision.

Worked example

payments.cards · card_number

4242 4242 4242 4242PAN· Restricted · 0.94

Regex flags the value as a candidate payment card. The model then reads the column name, the neighbouring fields and the surrounding records to decide whether this is a stored card number or a test fixture, a reference code, or padding.

Pattern only

Every 16-digit string is reported. The analyst triages the noise.

Hybrid

Context decides. Fewer findings to review, each with a confidence score.

Zero data egress

The model runs inside your perimeter.

With a local model configured, NLP inference happens on compute you allocate, on your own network, without sending scan data to an external AI service. Optional remote providers require explicit configuration and data-egress confirmation. Choose local inference when data must remain inside your perimeter.

Source access
Read-only
Internet for local inference
No

Screen quickly, or measure completely.

Two scan modes for two different questions. Quick scanning answers “is there personal data here at all?” in minimal time. Full scanning answers “how much, of what kind, and where exactly?”

Quick / Incremental

Screening scan

Full / Detail

Measurement scan

Primary objective

Rapid PII identification
Comprehensive PII discovery

Initial processing

Up to 10,000 rows
Entire dataset

Scan termination

Stops once PII is detected
Continues until completion

Output

Qualitative report
Quantitative report

Processing effort

Lower
Higher

Typical use case

Rapid assessment and screening
Baseline, compliance and detailed assessment

PII coverage

Early detection
Complete dataset coverage

Quick scanning evaluates up to 10,000 records from the selected source. If nothing is found it continues incrementally through subsequent records until PII is identified, then stops and reports.

The platform

One workflow, from connection to evidence.

Argon replaces fragmented, manual privacy processes with a single automated path — four steps your team runs, inside your own network, ending in something your team can review and prepare for an audit.

  1. 01

    Connect

    Point Argon at a database or file repository using read-only credentials over your internal network or VPC peering.

  2. 02

    Discover

    The engine interrogates schemas, tables and files to locate candidate personal data across the selected sources.

  3. 03

    Classify

    Findings are categorised by identifier type, sensitivity and the regulations that apply to them.

  4. 04

    Result

    You get a report with the affected tables and columns, masked sample values, confidence and rows impacted.

Everything the platform does.

Eight capabilities that take you from not knowing where personal data is, to having a monitored, governed and evidenced position on it.

AI-powered discovery

Scan databases and storage environments to identify and classify personally identifiable information automatically.

Scheduled scanning

Configure recurring scans so discovery stays current as data changes, rather than a one-off snapshot.

Masking and protection

Protect sensitive values through masking and controlled exposure to reduce unauthorised access and leakage.

Monitoring and alerts

Watch environments for changes, newly discovered sensitive data and emerging compliance risk, in real time.

Governance and classification

Centralised classification, ownership and policy controls to give the business visibility over its sensitive data.

Access rights management

Analyse user privileges, permissions and access patterns to support least-privilege on the data that matters.

Regulatory coverage

Map findings to the obligations that apply, across GDPR, the DPDP Act, CCPA/CPRA, POPIA, PCI DSS and LGPD.

Centralised risk view

A unified picture of sensitive data, exposure, access risk and compliance posture across the estate.

Compliance mapping

Findings, mapped to the clause that asks for them.

Discovery is only useful if it answers a specific obligation. Argon maps what it finds to the articles, sections and requirements your privacy and legal teams are actually working against.

Clause

Art. 5 — Principles

Discover PII, classify data, identify excessive or unnecessary data.

Art. 6 — Lawfulness

Identify data categories requiring processing-purpose and lawful-basis mapping.

Art. 9 — Special categories

Identify sensitive and special-category personal data.

Art. 25 — Protection by design

Support data discovery, minimisation and protection visibility.

Art. 30 — Records of processing

Support data inventory and processing-data mapping.

Art. 32 — Security of processing

Identify sensitive data exposure and support protection controls.

Art. 35 — DPIA

Support identification and assessment of high-risk personal-data processing.

Argon supports compliance activities through discovery, classification, risk assessment, regulatory mapping and evidence generation. It does not by itself guarantee regulatory compliance. Exact applicability depends on your jurisdiction, processing activities, data categories, contractual obligations and scope, and should be validated by your privacy, legal or compliance function.

Deployment

It runs where your data already is.

Argon installs into your environment rather than pulling data into someone else’s. Below is what the platform needs, so your infrastructure team can size it before the first conversation rather than after.

On-premises

Application and scanner deployed inside the corporate network, on hardware you control.

Private cloud / VPC

Deployed within your own private cloud or VPC environment, reachable over internal networking.

Air-gapped

No public internet access required when the native AI model is used — suitable for isolated environments.

Basic

Recommended

Deployment

On-premises / private cloud / VPC

Private network

Server

12 vCPU

24 vCPU

RAM

32 GB

128 GB

Storage

250 GB

1 TB + SSD

Operating system

RHEL 9.x / Ubuntu 22.04+

Latest supported enterprise OS

Network

1 Gbps

10 Gbps

IP address

Static IP recommended

Static IP

Hostname

Dedicated hostname

Dedicated hostname

DNS

Internal DNS resolution

Corporate DNS

Internet

Not required for local AI deployment

No public internet access

Before the first scan

  • Connectivity

    Network access to the target database and file repositories over VPC peering or a secure internal gateway.

  • Authentication

    Read-only database credentials for the discovery engine. Read/write only where the masking vault is integrated.

  • Compute

    Local compute allocated for LLM inference, so NLP parsing runs without data leaving the perimeter.

  • Configuration

    Scanning depth (sampling vs full-table) and the relevant pattern sets — GDPR, HIPAA, PCI DSS and others.

Actual sizing varies with data volume, number of sources, scan frequency, concurrent scans, file size and the AI model in use.

Who we are

Certified people behind the platform.

Netrik is a data security and privacy company specialising in PII scanning, data discovery and information risk management. Our mission is to make discovery and protection simple, accurate and actionable — without disrupting how the business runs.

Assessments and compliance consulting are delivered by certified practitioners. Every engagement is led by consultants holding globally recognised certifications, with hands-on experience across offensive security, governance, risk and compliance.

OSCPOffSecOffensive Security Certified Professional
CEHEC-CouncilCertified Ethical Hacker
CISMISACACertified Information Security Manager
CISSP(ISC)²Certified Information Systems Security Professional
CISAISACACertified Information Systems Auditor
CRISCISACACertified in Risk and Information Systems Control
ISO 27001ISMSLead Auditor / Lead Implementer

Our expertise areas

Information security

Control design, security architecture review and remediation planning for regulated environments.

Data privacy & governance

Discovery, classification, ownership and retention control over personal data at scale.

Regulatory compliance

DPDP Act, GDPR, ISO 27001, SOC 2, HIPAA and PCI DSS readiness and evidence generation.

Risk & audit management

Risk assessment, internal audit and third-party assurance across the supplier estate.

Answers to common questions

What security and compliance teams ask before a deployment. If something is not covered here, ask us directly.

Where is Argon deployed?

Either on-premises, with the application and scanner inside your corporate network, or in a private cloud or VPC environment you control. Both are supported.

What types of data sources can Argon scan?

Relational databases including MySQL, PostgreSQL and Microsoft SQL Server, plus Oracle and NoSQL databases; file servers and NAS; network shares; cloud databases; and object storage.

How does Argon detect and classify PII?

Three approaches. Regex does pattern-based detection of known formats. LLM does context-based identification of sensitive information. Hybrid combines both for improved coverage — pattern matching finds candidates, context analysis confirms them.

Does Argon modify or alter our data during scanning?

No. The scanning architecture uses read-only access to the data source. The scanner analyses data without requiring write access to the source database or repository.

Does our data leave our environment?

With a native or local AI deployment, scanning and AI-based classification occur within your controlled environment, without sending scan data to an external AI service. Optional external providers require explicit configuration and data-egress confirmation. Your selected deployment and provider determine where inference runs.

Is our data used to train the Argon AI model?

No. Customer scan data is not used to train, fine-tune or improve the underlying AI model unless explicitly authorised under the applicable agreement and product configuration.

Are scan sessions isolated between customers?

Yes. Complete separation is maintained between customers, with no cross-customer data sharing or retention. With a local deployment, we do not have access to your data at all.

Can Argon help with privacy and regulatory compliance?

Yes. Argon supports compliance activities by mapping discovered and classified PII to applicable requirements under GDPR, the India DPDP Act, CCPA/CPRA, POPIA, PCI DSS and LGPD. It supports those activities; it does not by itself guarantee compliance.

What technical support is available?

Technical support covers product installation, configuration, connectivity, scanning and troubleshooting, according to your support plan.

How do we raise a support request?

Through the designated support email, the support portal, or another approved channel agreed during onboarding. A ticket reference is generated so the request can be tracked.

Get started

See what a scan finds in your own environment.

The most useful version of this conversation is the one run against a representative sample of your data, not a generic demo dataset.

Email the team

  1. 01Talk to a solutions engineer

    We come back within one business day to understand your environment and priorities.

  2. 02A walkthrough on your data

    A 30–45 minute demonstration scoped to a representative sample of your estate.

  3. 03Findings you can take upstairs

    A summary of key risks, affected systems and the regulatory obligations they touch.