Buyer's guide

Best PII masking tools for test data in 2026

PII masking tools find personal data in the systems you copy into test, replace it with realistic values and keep those values consistent so tests still work. Ten tools compared on discovery, consistency across systems, documents, synthetic data and where they run. Synthesized publishes this guide and is one of the tools listed; claims about other vendors are sourced and dated at the end of the page.

PII masking tools at a glance

Synthesized first, then alphabetical. Facts as stated on each vendor's own pages and documentation in November 2026; "not stated" means we could not find it there.

Finds PIISame values across systemsDocuments and free textSynthetic dataWhere it runs
SynthesizedYes, detects sensitive fields and applies configured policiesYes, one replacement value per person across every systemYes: PDFs, scans, email, chats, transcriptsYesOn-premises, including air-gapped, or SaaS
Broadcom Test Data ManagerYes, profiling and PII discoveryNot statedNot statedYesOn-premises
DATPROFYes (Analyze)Not statedFile masking (xPrivacy)Yes (Privacy)On-premises; Runtime on AWS (bring your own licence)
Delphix (Perforce)Yes, profiling of columns and dataYes, consistent across sourcesStructured files only (delimited, XML, JSON, Parquet)Early AccessCustomer-managed on most clouds and hypervisors
GenRocketNot statedNot statedNot statedYes, synthetic firstDesign in SaaS; generation engine on-premises
IBM OptimNot statedNot statedNot statedIBM's synthetic data product is in maintenance modeOn-premises, including z/OS
Informatica TDM (Salesforce)Yes, sensitive data discoveryNot statedNot statedRule-based generationOn-premises
K2viewYes, automated discovery and classificationYes, in flight by business entityThrough a third-party partner, per K2viewYesSaaS, self-hosted or air-gapped
TonicYes, Structural sensitivity scanYes, consistency across databasesYes, Tonic Textual (sold separately)Yes, Tonic FabricateTonic Cloud; self-hosted on Enterprise plans
XDM (UBS Hainer)Yes, PII finderYes, consistent across platformsNot statedYesMainframe, distributed and cloud

If something about a vendor has changed, tell us and we will update this page.

What to check before you choose

Six questions separate PII masking tools. Run each one on a copy of your own data, with your own edge cases, rather than a demo schema.

Does it find the PII you didn't know about?

Personal data hides in free-text notes, comment fields, JSON columns and custom tables. Ask how discovery works (column names, data sampling or both) and what it reports when it is unsure.

Is the same person masked the same way everywhere?

A customer masked one way in the core system and another way in the CRM breaks end-to-end tests. Ask for deterministic masking across every database, file and refresh in scope.

Does masked data still behave like real data?

Names should fit their countries, addresses should be real places, and IDs, IBANs and card numbers should keep valid formats and check digits, or tests fail for the wrong reason.

What about documents, emails and logs?

PDFs, scans, emails, chat logs and application logs often hold the same personal data as the tables. Check whether the tool masks them under the same policy, or whether you need a second product.

Can you prove what was masked?

Auditors and regulators ask for evidence. Look for a report of what each run detected and masked, not just a list of algorithms.

Where does it run?

If production data cannot leave your network, check whether the tool runs on-premises or air-gapped, and whether any data goes to a vendor's SaaS.

The tools, one by one

Synthesized

Detects sensitive fields, then masks, subsets and generates test data across databases and applications such as SAP S/4HANA, Salesforce, Oracle, Temenos and Avaloq, including mainframe Db2 for z/OS 13 and later, Db2 for i and VSAM, with one replacement value per person across every system. Documents, emails, chats and call transcripts are masked under the same policy as the tables. It runs on-premises, including air-gapped sites, or as SaaS, and delivers data through an API, CI/CD pipelines, Tricentis Tosca and UiPath. At Deutsche Bank, the Credit Risk Technology team cut test data provisioning time by 80% (case study).

Watch for: there is no free tier (a free trial is on AWS Marketplace), and document support is delivered with Synthesized engineers today.

Broadcom Test Data Manager

Formerly CA Test Data Manager. Covers profiling and PII discovery, masking, subsetting, synthetic generation, a self-service portal, virtual copies and a test data warehouse, including mainframe Db2, IMS, IDMS and VSAM.

Watch for: document masking is not stated, and pricing is not public.

DATPROF

Separate products for masking and synthetic data (Privacy), subsetting (Subset), self-service and API delivery (Runtime), discovery (Analyze) and virtualization (Virtualize), plus xPrivacy for masking files. Licences are priced per module rather than by database size, and there is a 14-day free trial.

Watch for: SAP support is not stated.

Delphix (Perforce)

Profiles data sources to find sensitive columns, by column name and by scanning the data, then assigns masking algorithms by domain so the same values are applied consistently across sources. It masks structured files as well as databases, and delivers masked data as full-size virtual copies. See Synthesized vs Delphix.

Watch for: its newer automated discovery profiler is limited to specific databases, synthetic data is Early Access, and free-text documents are not stated.

GenRocket

Synthetic data first. It replaces sensitive values with synthetic ones, including Db2 z/OS and IMS data and EBCDIC files described by COBOL copybooks. Test data is designed in a SaaS app, while the generation engine runs on your premises and only metadata leaves.

Watch for: PII discovery is not stated, and it suits generated data more than masked copies of production.

IBM Optim Test Data Management

Provisions complete business objects with relationships intact, right-sizes data and applies policy-based masking, with deep z/OS roots: Db2 z/OS, IMS, VSAM and sequential files.

Watch for: the latest release we could confirm is 11.7.0.2 (November 2023), and IBM Client Engineering says its synthetic data product, Test Data Fabrication, is in maintenance mode.

Informatica Test Data Management

Now part of Salesforce, which completed its acquisition of Informatica on 18 November 2025. Offers sensitive data discovery, subsetting, masking, rule-based generation, a self-service portal and a test data warehouse.

Watch for: document masking is not stated.

K2view

Discovers and classifies PII automatically, using regular expressions, rule sets and generative AI, then masks it in flight as it organises data by business entity, so a customer is masked the same way across systems. Runs as SaaS, self-hosted or air-gapped. See Synthesized vs K2view.

Watch for: its own Tonic comparison says documents go through a third-party partner, and its AWS Marketplace listing starts at $120,000 a year.

Tonic

Three products sold separately: Structural for databases, with a sensitivity scan that flags sensitive columns and recommends generators, and consistent values across databases; Textual for documents and audio, which detects entities in 50+ languages; and Fabricate for generating data from scratch. See Synthesized vs Tonic.

Watch for: databases and documents are separate products, and self-hosting needs an Enterprise plan.

XDM (UBS Hainer)

A German TDM platform covering masking, subsetting, synthetic generation and provisioning. A PII finder scans databases for sensitive data, and masking stays consistent across platforms, from cloud storage and relational databases to VSAM.

Watch for: document masking and pricing are not stated on its pages.

Questions about PII masking for test data

What are the best PII masking tools for test data?

It depends on your estate. For mainframe-heavy estates: Broadcom, Delphix, DATPROF, GenRocket, IBM Optim, Synthesized and XDM. For full-size virtual copies: Delphix. For self-serve document redaction: Tonic Textual. For databases, SAP and core applications with documents, emails and transcripts masked under the same policy, on-premises or as SaaS: Synthesized. Shortlist two or three and test them on your own data.

What is the difference between PII masking and synthetic data?

Masking keeps real records and replaces the personal values in them, so data keeps its real shape and history. Synthetic data creates new records that follow the patterns of real data, with no real person underneath. Many teams mask production for regression tests and generate data for new features and edge cases.

Is masked test data still personal data under GDPR?

It can be. If masked records can still be linked back to a person, for example through untouched fields or a reversible mapping, they are pseudonymised and remain personal data. Data counts as anonymised only when people can no longer be identified. Ask each vendor which of its methods are reversible, keep any mapping keys out of test, and check with your data protection officer how your masking is classified.

Can PII masking tools mask documents and logs as well as databases?

Some can. Synthesized masks PDFs, scans, emails, chats and call transcripts under the same policy as tables, and Tonic Textual redacts or synthesizes documents self-serve. Delphix masks structured files such as CSV, XML and JSON, and DATPROF masks files with xPrivacy. For most other tools on this list, document masking is not stated. See the best unstructured data masking tools.

See it on your own data

Book a demo for your databases and applications, or prove document masking on 50 of your own files in 48 hours.

Updated November 2026

All product names are trademarks of their respective owners. This guide is based on public information as of November 2026 and is not legal advice.