Data residency for testing comes down to three questions: where does your test data physically live, what kind of data is in it, and who can reach it? Most UAT teams only answer the first one, and only for the test environment. The recordings, screenshots, transcripts and tickets your testers create are test data too. This guide shows how to map all of it, decide what belongs where, and choose tooling that makes the answers easy to document. Wezard captures UAT issues as screen recordings with real-time transcription and sends them to Jira, Azure DevOps or ServiceNow, so the bug-reporting path is one you can map once and keep audit-ready.

Wezardapp
Make UAT Issues Easier to Map
Wezard records UAT issues, transcribes them in real time, and detects duplicate tickets before they reach Jira or Azure DevOps.

Visit

TL;DR

  • Data residency is where data is stored and processed. Data sovereignty is which country’s laws govern it. They are related but not the same.
  • The biggest residency risk in UAT is usually production data copied into test systems, not the test tool itself.
  • Use synthetic data first. If you must use production data, mask it and apply production-like controls.
  • Map every place test data travels: environment, recordings, attachments, transcripts, backups, integrations and subprocessors.
  • Document your decisions in a data inventory and data-flow diagram so an auditor can follow them.

Table of Contents

Residency, sovereignty and localization in plain terms

These three terms get mixed up in vendor questionnaires, so settle them first.

  • Data residency is the geographic location of the systems that store or handle the data.
  • Data sovereignty is the legal authority over that data and the laws that apply to it.
  • Data localization is a legal requirement to keep data where it was created.

That is how IBM’s explainer separates them. It also notes that localization rules vary: some require an in-country copy, others prohibit transfers abroad.

For a UAT team, the practical meaning is simple. A bug recording stored in a given region is a residency fact. The law that applies to that recording is a sovereignty fact. Your compliance or legal contact decides which rules matter. Your job is to supply accurate facts about where things sit.

Why testing is where residency slips

Production systems usually get careful review. Test environments often do not. The FFIEC’s IT Examination Handbook, written for US financial regulators, defines non-production environments as development, testing and quality assurance. It warns that copying production data into them can expose information to people who have no job need for it, and that security may be weaker there.

Tester running an acceptance test in a QA lab

Thoughtworks makes the same point in its technology guidance on production data in test environments. Test systems often have weaker privacy protection than production, and masking only some fields can leave other data exposed. It adds that copying production data to test systems hosted or accessed from another country or region can create privacy-law risk, especially in complex cloud setups.

Here is what that looks like on a real UAT cycle:

  • A business user tests a claims screen with a real customer’s record on it.
  • They record their screen and narrate the problem.
  • The recording, the transcript and the ticket now hold that customer’s details, in places your data inventory never listed.

The residency question is no longer only about the test database. It covers every artifact the testing produces.

What counts as test data in UAT

Teams asking for examples of testing data usually think of database rows. For UAT, widen the list:

Artifact Typical content Residency question
Test database Records, accounts, transactions Which region and which environment?
Screen recordings Anything visible on screen Where is video stored and backed up?
Voice and transcripts Narration, names said aloud Where is audio processed?
Screenshots and attachments Files, exports, documents Who can download them?
Tickets in Jira, Azure DevOps or ServiceNow Descriptions, links, logs Which instance, which region?
Backups and logs Copies of all of the above How long are they kept, and where?

If an item is on this table and not in your data inventory, you have found a gap.

Choose the data before you choose the region

The cleanest way to solve a residency problem is to have less regulated data to place. The order of preference from official guidance is clear.

  1. Synthetic data. The FFIEC handbook recommends keeping non-production environments independent of production and using simulated synthetic data whenever possible. The International Foundation for Information Technology describes UAT data that is realistic and complete enough for acceptance testing but generated to represent usage patterns, without real individual records.
  2. Masked or sanitized production data. The same handbook recommends masking or sanitizing production data whenever feasible. Oracle’s documentation explains that masking and subsetting reduce exposure, and that copying production datasets widens your sensitive-data security and compliance boundary.
  3. Production data with production-like controls. If neither works, the handbook calls for approval and controls similar to production: identity and access management, password controls, logging and monitoring.

A business analyst can drive this. Put the data choice in your UAT entry criteria, so testing does not start until someone has signed off on what data the testers will see. Our guide to acceptance testing entry criteria covers where that sign-off fits, and the UAT environment requirements checklist lists what to verify in the environment itself.

Map where test data goes

Governance starts with an inventory. The FFIEC’s data guidance says to record each dataset’s storage location, criticality and sensitivity, and to keep controls consistent with those characteristics. It cites data-flow diagrams, entity-relationship diagrams, database schemas and data dictionaries as ways to document how a system uses data.

For a UAT cycle, a one-page version works:

  1. List every data type your testers will touch: customer records, employee data, payment data, health data.
  2. Rate each one for sensitivity: public, internal, confidential, regulated.
  3. Draw the flow. Start at the tester’s screen and follow the data: capture, upload, storage, processing (such as transcription), integration into your tracker, backup.
  4. Write the location next to each stop: region, provider, and who administers it.
  5. Record the decision for each: allowed, allowed with masking, or not allowed in test.

AWS’s Digital Sovereignty Lens suggests the same approach at larger scale: map residency requirements by data type to permitted storage, processing and transfer locations, and consult legal and privacy experts where needed. For documentation habits that survive an audit, see our overview of software testing documentation types and what a test audit involves.

A region setting alone does not prove compliance. Region choice is one control among access, encryption, monitoring and vendor management, as this cloud data compliance guide explains.

Questions to answer about any UAT tool

When you evaluate UAT tools, ask in this order. Each question maps to a row of your data-flow diagram.

  • Storage: Where are recordings, screenshots and attachments stored, and in which country or region?
  • Processing: Where does transcription and any AI analysis run?
  • Backups: Where do copies sit, and for how long?
  • Access: Who at the vendor can view customer content, and under what controls? Is single sign-on supported?
  • Integrations: Which trackers receive the data, and are those instances in a region you have approved?
  • Subprocessors: Which third parties handle data on the vendor’s behalf, and does a data processing agreement cover them?
  • Retention and deletion: How long is test content kept, and can you remove it?

The vendor chain matters as much as the storage region. WilmerHale’s review of 2025 US privacy developments notes that the Department of Justice bulk-data rule, which took effect on April 8, 2025, restricts certain transactions involving bulk sensitive US personal data and countries of concern or covered persons. Its practical lesson for you: review onward transfers and who sits in the vendor chain, not only the region name on a settings page. Whether the rule applies to your organization is a question for your counsel.

Once you have the answers, verify them. AWS’s guidance supports documenting where data is stored, processed and transferred, then validating in a test account that controls block prohibited actions, and monitoring continuously afterward.

Keep the capture itself clean

Most residency exposure in UAT comes from what testers capture without thinking. Three habits reduce it:

Sharing is the next leak point. Recording links that travel by email or chat leave your mapped flow, so read how to share bug recording links safely. For protection in storage and transit, see encryption for bug reports. If your UAT touches patient data, the three HIPAA bug-reporting controls apply, and our test data privacy guide and video privacy guide go deeper.

A worked example

Imagine a regional insurer running UAT on a new claims portal with 12 business users.

  1. The BA lists data types: claimant names, policy numbers, medical notes. Medical notes are rated regulated.
  2. The team agrees on synthetic claimants for all UAT scripts. Medical notes are fully fictional.
  3. The PM draws the flow: tester screen, recording, transcript, Jira ticket, backup. Each stop gets a region and an owner.
  4. The QA lead confirms that the tooling’s data processing agreement covers the recording and transcription steps.
  5. The decision log is stored with the UAT plan, ready for the next audit.

Total effort is a couple of hours of planning. The payoff is that when someone asks where claimant data went during testing, you can answer from a document instead of from memory.

Our perspective

Residency is not a setting you flip. It is a map you keep. In UAT, the map has to include what testers create while they work: recordings, narration, transcripts and tickets.

We built Wezard so that the reporting path is short and visible. A tester records the screen, the issue is transcribed in real time, and AI checks for duplicates before the ticket reaches Jira, Azure DevOps or ServiceNow. Fewer hops and fewer duplicate copies mean fewer places to document. Our platform offers enterprise-grade security with SSO, we hold SOC 2 certification, and our data processing agreement is published for your legal team to review.

Start from the data flow, then pick tooling that fits it, not the other way around.

Marketing

Where Wezard fits

Wezard is user acceptance testing software for business users, analysts, project managers and QA teams. Testers record the issue and talk through it, Wezard transcribes it as they go, and its AI flags likely duplicates before they reach your backlog. Plans are published: Starter at $25 per month, Team at $70 per month and Business at $160 per month, with Enterprise priced on request. You can see them on the pricing page, and the use cases page shows how teams run UAT with it.

Wezardapp

FAQ

What is the difference between data residency and data sovereignty?

Residency is where data is physically stored and processed. Sovereignty is whose laws have authority over it. Data can sit in one country and still be subject to another country’s rules, which is why both belong in your documentation.

Do I need to keep UAT data in the United States?

Not necessarily. The US has no single national residency rule for test data, but sector guidance, contracts and rules such as the Department of Justice bulk-data rule can limit where certain data may go. Ask your compliance or legal team which apply, and record the answer in your data inventory.

Is synthetic data enough for acceptance testing?

Often yes. Official guidance recommends synthetic data whenever possible, and it can be built to mirror realistic usage patterns. Where a scenario truly needs production-derived data, mask it and apply production-like access controls and logging.

What should a data processing agreement cover for a UAT tool?

It should state what customer content the vendor handles, how it is protected, which subprocessors are involved and how data is deleted. Have your legal or privacy contact review it before testing begins.

Sources

Share This Story, Choose Your Platform!