AI processing stays in the EU

For enterprise buyers, data residency isn't optional. It's a procurement requirement. Where does the data go? Who processes it? Under what terms?

Clarity First routes all AI processing through Microsoft Azure OpenAI in Sweden Central (EU). Interview data is stored in Supabase (Frankfurt, EU). No interview content leaves the European Union for AI processing.

This isn't a future roadmap item. It's how production works today.

The architecture

Here's what happens when a stakeholder completes an interview:

1. Interview capture

The conversation — voice or text — is captured and stored in our database (Supabase, Frankfurt, EU). Voice is transcribed in real-time during the interview.

2. AI analysis

When analysis runs, the interview transcripts are sent to Azure OpenAI (Sweden Central, EU) for processing. Each of our eleven AI agents — Humboldt, Bayes, Linnaeus, Darwin, Curie, Popper, and the rest — makes API calls to the Azure endpoint. The processing happens within Microsoft's EU infrastructure.

3. Results storage

Analysis outputs (findings, patterns, recommendations, evidence references) are stored back in our database in Frankfurt. The complete evidence chain — from original quote to final recommendation — lives in the EU.

What about sub-processors?

Our Data Processing Agreement lists every sub-processor that handles personal data:

| Sub-processor | Purpose | Location | |---|---|---| | Microsoft (Azure OpenAI) | AI text processing | Sweden Central, EU | | Microsoft (Graph API) | Teams calendar & transcript fetch | EU | | Supabase | Database & authentication | Frankfurt, EU | | Vercel | Application hosting | EU edge nodes | | ElevenLabs | Text-to-speech for AI interviewer | US (SCCs in place) | | Resend | Transactional email | US (SCCs in place) | | Tavily | External research (Curie agent) | US (SCCs in place) | | Stripe | Payment processing | US/EU (SCCs in place) |

The core processing pipeline — interviews, analysis, storage — stays entirely within the EU. Only auxiliary services (voice synthesis, email delivery, external research, payments) involve US-based processors, and these operate under Standard Contractual Clauses.

No training on your data

Azure OpenAI operates under Microsoft's enterprise terms. Your interview data is:

Not used to train models — Microsoft's Azure OpenAI service does not use customer data for model training
Not shared — Your data is isolated to your processing requests
Not retained — Azure OpenAI does not store prompts or completions beyond the immediate API call

This is a key difference from consumer AI services. Enterprise Azure OpenAI provides the model capabilities without the data sharing.

Microsoft DPA coverage

Microsoft's Azure services are covered by their Data Processing Addendum (aka.ms/dpa), which includes:

GDPR compliance commitments
Standard Contractual Clauses for international transfers
Technical and organisational security measures
Audit rights and transparency obligations

This sits alongside our own DPA and Privacy Policy, which detail how Clarity First handles data end-to-end.

What this means for procurement

If you're evaluating Clarity First for enterprise use, here's what you can tell your procurement and compliance teams:

AI processing: Azure OpenAI, Sweden Central, EU — no interview data leaves the EU for AI processing
Data storage: Supabase, Frankfurt, EU — all interview and analysis data stored in the EU
Data ownership: You retain full ownership. Data can be deleted at any time.
Model training: Your data is never used to train AI models
Compliance: GDPR compliant. DPA available. Sub-processor list published and maintained.
Audit trail: Full evidence chain from interview quote to final recommendation

Questions?

If your compliance team needs additional detail, we're happy to walk through the architecture. Book a call or email us at hello@clarityfirst.io.

For full legal documentation, see our Privacy Policy, Terms of Service, and Data Processing Agreement.

EU Data Processing: How Clarity First Handles Your Data