Security & data

How we handle your data.

You’re trusting us with customer conversations, internal documents and business processes. Here’s exactly how we protect them, in plain language.

SingaporeClient data hosted in Singapore by default
ZeroClient data used to train AI models
30 daysTo delete our copies after handover
72 hoursTo tell you about any incident affecting your data

Where your data lives

Hosted in Singapore, by default.

Anything we host for you runs in the Singapore region of a major cloud provider: Amazon Web Services, Google Cloud or Microsoft Azure. Prefer your own cloud account or servers? We build there instead, so your data never leaves infrastructure you control.

Singapore region first

Databases, files, self-hosted workflows and backups for client projects stay in Singapore unless you agree otherwise in writing.

Your infrastructure, if you prefer

Self-hosted n8n and custom code can run entirely inside your own cloud account or data centre, under your own security policies.

Nothing moves without telling you

If a service you need only runs overseas, such as a particular AI model, we tell you before we build, and the transfer is covered by contract terms that keep the data protected to a comparable standard, as the PDPA requires.

AI providers

Your data never trains anyone’s AI.

We use the business and API services of established AI providers, such as OpenAI, Anthropic and Google, whose commercial terms exclude customer data from model training by default. We never use your data to train shared or public models, including our own.

You know which model

Before we build, we tell you in writing which AI provider and model each automation uses, and where it processes data.

Minimum data, every time

Automations send a model only what the task needs. Where possible, names, IDs and contact details are removed or masked before anything leaves your systems.

Shortest retention available

We switch on the shortest data-retention settings each provider offers, including zero-retention options where they’re available.

Private models for sensitive data

For confidential or regulated data, we can use a private model trained or hosted for you alone, so it never touches a shared service.

How we protect it

Ten controls on every project.

Encryption

In transit with TLS 1.2 or higher, and at rest with AES-256 on every database, file store and backup we manage.

Only what the job needs

Each automation can reach only the specific data and actions its one job needs. Nothing gets a broader export “just in case”.

Credentials kept out of automations

Keys and passwords live in an encrypted secrets vault and are supplied only when an action runs, so revoking one cuts access everywhere.

Multi-factor sign-in

Every account our team uses to reach client systems is protected by multi-factor authentication and a password manager.

Need-to-know access

Only the people working on your project can reach your data. We review access every quarter and remove it the day someone leaves a project.

Checks on the server

Every permission is enforced on the server, not just hidden behind a button, so access can’t be bypassed.

A person in the loop

New automations run supervised: actions wait for your team’s OK until you’re confident they can run on their own.

Everything logged

Every run is recorded with when it ran, what triggered it and exactly what it did.

You can stop it

A kill switch your team controls, quick undo on reversible actions, and instant rollback to the last working version.

Updates and backups

We keep the systems we manage patched, and back them up automatically so we can restore quickly.

The four-phase safety process behind every automation

Retention

How long we keep it.

We keep personal and business data only as long as the project needs it. On request, we confirm deletion in writing.

WhatCoversHow long
During the projectOnly the data each automation needs, in the Singapore-hosted environment for your project, or in your own systems.Until handover
Copies we hold after handoverSample files, exports and test data used while building.Deleted within 30 days of handover
Encrypted backupsPoint-in-time backups of anything we host for you.Expire automatically within 90 days
Activity logsRecords of what each automation did and when.12 months by default, or as you choose
Anything on your infrastructureWorkflows, code and data on servers or accounts you own.Yours: kept or removed on your instructions

If something goes wrong

Our incident response.

No system is risk-free. What matters is how fast and how openly a problem is handled.

  1. Contain. We pause the affected automation with its kill switch and revoke any credentials involved, straight away.
  2. Tell you. We notify your named contact within 72 hours of confirming an incident that affects your data, with what we know so far.
  3. Investigate. We use the activity logs to establish exactly what happened, what data was involved and who was affected.
  4. Support your obligations. If it’s a notifiable data breach under the PDPA, we give you what you need to notify the PDPC within 3 calendar days, and the people affected where required.
  5. Fix and report. We fix the cause, and send you a written report of what happened and what we changed.

Standards

Built around Singapore’s rules.

Our practices follow Singapore’s Personal Data Protection Act (PDPA), the PDPC’s advisory guidelines on using personal data in AI systems, and IMDA’s Model AI Governance Framework.

Personal Data Protection Act

We design automations to collect only what a task needs, use it only for that purpose, protect it, and keep it only as long as it’s needed.

A named data protection contact

Questions about how we handle personal data go to help@wunderwaffen.co, and reach the person responsible for data protection at WunderWaffen.

Honest about certifications

We don’t display certifications we don’t hold. If your procurement team needs evidence against ISO 27001 or SOC 2 controls, we’ll work through your questionnaire control by control.

AI automation and the PDPA: a practical checklist

For security teams

What you can ask us for.

Send your questionnaire or requests to help@wunderwaffen.co, or book a call and talk it through.

  • A non-disclosure agreement, before you share anything
  • A data processing agreement
  • Your completed security questionnaire
  • A list of the AI providers and services your project uses
  • Hosting inside your own cloud account
  • Written confirmation of deletion when we’re done

Related: Privacy Policy · Cookie Policy · Website Terms

Questions

Security questions

Do you sign NDAs?

Yes. Many clients ask for one before sharing details of their processes, and we’re happy to sign before the discovery call.

Who on your team can see our data?

Only the people working on your project, and only what the project needs. Automations themselves get the same least-privilege access: just the data their one job requires.

Can automations run on our own infrastructure?

Yes. Self-hosted workflows and custom code can run on servers or cloud accounts you control, so your data stays where your policies say it should.

What happens to our data when a project ends?

We delete the copies we hold, such as sample files, exports and test data, within 30 days of handover, and encrypted backups expire within 90 days. We confirm deletion in writing on request. Anything built on your own infrastructure simply stays with you.

Is our data used to train AI models?

No. We use the business and API services of AI providers whose commercial terms exclude customer data from training by default, and we never use your data to train shared or public models, including our own.

Where is our data hosted?

In the Singapore region of a major cloud provider by default, or inside your own cloud account if you prefer. If a service you need processes data overseas, we tell you before we build.

Will you tell us if something goes wrong?

Yes. We notify your named contact within 72 hours of confirming an incident that affects your data, and support you with any PDPA breach notification you need to make.

Can you complete our vendor security questionnaire?

Yes. Send it to help@wunderwaffen.co.

Free consultation

Get 30 minutes with our experts.

Explore how AI can automate your most complex workflows. No obligation, just a clear plan for what to automate first.