We use AI tools configured for privacy, hosted on EU infrastructure. Here's exactly how we handle your data.
Consumer AI products and business API access have different data policies. We use the business-grade approach.
May use conversations to improve models (opt-out available). Data retained longer. Fewer contractual protections.
No training on your data. Contractually prohibited. 30-day retention max. Full DPA available.
We use Claude with "Improve Claude" disabled. We never use thumbs up/down feedback on client conversations. Training is opt-in only, and we've opted out.
Hosted solutions for clients run on EU-based infrastructure.
How we handle data during AI transformation workshops.
| Stage | What Happens | Retention |
|---|---|---|
| During workshop | Data processed via Claude API (no training) | 30 days at Anthropic |
| Deliverables | Final outputs delivered to you | You control |
| Our systems | Working files on encrypted drives | Deleted after project |
For clients requiring data to never leave the EU, we offer AWS Bedrock EU deployment.
Claude models running in AWS Frankfurt (eu-central-1), Ireland (eu-west-1), or Paris (eu-west-3). Data never leaves EU jurisdiction. You control all logs via CloudWatch.
| Aspect | Standard (Claude API) | Bedrock EU |
|---|---|---|
| Data location | US (with SCCs) | EU only |
| Training | Never (API) | Never |
| Logs | Anthropic systems | Your AWS account |
| Legal basis | Anthropic DPA + SCCs | AWS DPA (EU) |