Company logo|Trust Center

Thoughtful Oasis

Thoughtful Oasis is an AI research lab in Enschede, the Netherlands, building interpretable systems for mission-critical work in government and industry, operating under an ISO/IEC 27001:2022-certified information security management system.

Compliance

Industry standards and compliance frameworks this organisation is aligned with or currently certified against.

GDPR:2020
Certified: ISO27001:2022

Resources

Downloadable security policies, compliance reports, and certificates. Some documents may require an access request.

Security Documentation

Statement of Applicability (SOA) - ISO27001:2022 - 2025

Our Statement of Applicability for ISO27001:2022

Certifications

ISO/IEC 27001:2022 certificate

Our ISO 27001 certificate.

Controls

Specific security measures this organisation has implemented, alongside their current operational status.

General EffectiveAcceptable use policy establishedEffectiveAccess and devices revoked when employees leave and contracts endEffectiveSecure software development policy established View 6 more
Operations Management EffectiveChange management policy establishedNot verified as effectiveCompliance monitoredView details
Data Security & Privacy EffectiveData leakage prevention (DLP) technology in useEffectiveData protection impact assessments (DPIAs) performedEffectiveEncryption of data at rest View 7 more
Information Security Governance EffectiveDocument control system implementedEffectiveInformation security objectives definedEffectiveInformation security policy established View 6 more
Threats & Vulnerabilities EffectiveAnti-malware technology appliedEffectivePenetration testing performedEffectiveThreat and vulnerability information collected and followed up View 3 more
Business Continuity EffectiveBackups performed and redundancy implementedEffectiveBusiness impact analysis performedEffectiveBusiness continuity plan established View 1 more
Logging & Monitoring EffectiveLog files securedEffectiveResource capacity monitoredEffectiveLog files created for security events View 1 more
Access Control EffectiveAccess roles and rights definedEffectiveAccess rights reviewedEffectiveAccess control policy established View 1 more
Third-Party Management EffectiveVendor security policy establishedNot verified as effectiveVendors monitoredView details
Product & Service Development EffectiveSecure coding practices appliedEffectiveSoftware tested before being put in productionEffectiveDevelopment, test and production environments separated View 3 more
Fair Operating Practices EffectiveOperating procedures establishedView details
Asset Management EffectiveAsset inventory maintainedView details
People & Culture EffectiveInformation security in employment contractsEffectiveScreening performed for employeesEffectiveEmployees trained in line with job responsibilitiesView details
Incident & Problem Management EffectiveCommunication structure establishedEffectiveIncidents followed up in line with the incident response planEffectiveIncidents evaluated and recordedView details
Compliance & Legal EffectiveLegal and contractual requirements identifiedEffectiveInternal audits performedEffectiveCorrective actions takenView details
System & Network Security EffectiveUser endpoint devices managedEffectiveNetwork segmentation implementedView details
Risk Management EffectiveVendor agreements reviewed and registeredEffectiveRisk assessments performedEffectiveRisk treatment definedView details

Subprocessors

Third parties this organisation works with that may process customer data.

MoneyBird Internal Operations
NL

Online bookkeeping software for TO finance.

Mistral AI & Machine Learning
FR

LLM provider used for document summarisation and inference.

LLM Whisperer AI & Machine Learning
US

PDF parsing API used to extract structured text from documents.

Runpod Cloud Infrastructure
US

GPU compute provider used for occasional model inference workloads.

Sentry Analytics & Monitoring
US

Error tracking and application monitoring for our production apps.

Frequently asked questions

Answers to commonly asked questions regarding this organisation's security and privacy practices.

Access & Authentication

How is internal access to systems and customer data authenticated?

Staff sign in through a single Google Workspace identity, with single sign-on enabled wherever a system supports it. Multi-factor authentication is required for all administrative and production access.

Compliance & Certifications

Are you ISO 27001 certified?

Yes, our information security management system is certified against ISO/IEC 27001:2022. Our Statement of Applicability is available in this Trust Center on request.

Data Protection & Encryption

Are you subject to NIS2 (the Dutch Cyberbeveiligingswet), and how do you support our NIS2 supply-chain obligations?

Thoughtful Oasis is not designated as an essential entity under the Cyberbeveiligingswet (NIS2). As a result we have no own registration obligation on MijnNCSC. As a supplier to organisations that do fall under NIS2, we support your supply-chain obligations through an information security management system that is certified against ISO/IEC 27001:2022 and aligned with the NIS2 Quality Mark (Basic). The five basic principles that the NCSC sets out for suppliers are covered: 1. Identifying risks, 2. Promoting secure behaviour, 3. Protecting systems, applications and devices, 4. Managing access, 5. Preparing for incidents. Security incidents are handled under our Incident Response Framework, and we support customers in meeting their own significant-incident notification duties.

How are backups handled?

The primary database is backed up daily, using the platforms' native backup and versioning rather than bespoke pipelines. Backups are verified and disaster recovery is tested periodically.

How are secrets and credentials managed?

Application secrets are held in Heroku config vars, GitHub Actions secrets, or LastPass, never hard-coded. Code reviews check for hard-coded secrets, and any secret accidentally committed is rotated immediately.

Is my data encrypted?

Yes, data is encrypted in transit with Transport Layer Security and at rest by the managed platforms that store it, such as Heroku Postgres. All employee devices use full-disk encryption.

EU AI Act

What does "auditable by construction" mean, and can we retrieve the audit trace for a specific decision?

Every input is encoded into a vector ontology with named, defined, versioned axes, so there is no opaque embedding step. For any classification, retrieval or recommendation we can show the audit trace: the named dimensions that drove the result and the ontology version live at that moment.

Hosting & Data Residency

Is customer data processed outside the EU?

The primary data plane sits in EU regions. A number of subprocessors, for example source-code hosting and some AI model providers, are US-based, with transfers covered by EU adequacy decisions where available and otherwise by Standard Contractual Clauses; the full list is in the Subprocessors section.

Where is your service hosted and where is customer data stored?

Our primary data stores sit in EU regions: Heroku's EU data plane, AWS in Luxembourg and Ireland, background compute on Hetzner in Germany, and the vector database in the Netherlands. Some platform administration and subprocessors are US-based, covered by EU adequacy decisions or Standard Contractual Clauses.

Incident response

How do you handle security incidents and vulnerability reports?

Report a suspected incident or vulnerability to privacy@thoughtful-oasis.com, monitored on business days. One coordinator owns it end to end: we acknowledge inside 24 hours, contain inside 72, and notify affected customers in line with their obligations. Every confirmed incident closes with a written summary: what happened, what was affected, and what changed.

Privacy & GDPR

Are you GDPR compliant?

Yes, we process personal data in line with the GDPR and our Privacy Policy. We act as a data processor, with the customer as controller.

Secure Development & Vulnerability Management

How do you manage vulnerabilities?

Dependencies, source code, and secrets are scanned continuously with Dependabot, GitHub CodeQL, and secret-scanning across all active repositories. We monitor vendor and community advisories, and track runtime errors in Sentry.