Services Secure AI Workspace

Private AI for your company, inside your own AWS account in the EU.

Your employees already use AI, often through personal accounts with company documents and no record. Secure AI Workspace gives them a private Claude assistant that runs in your AWS account, so you control who uses it, what data it sees and where that data goes, and you can prove it to your auditor.
Full visibility
Every request is logged in your own AWS account: who asked, which model answered, and when. You always know what your company sends to AI.
EU data residency
Requests are processed by Claude on Amazon Bedrock in EU regions, and each log records where the data was processed.
No training on your data
Your prompts and documents are used only to answer the question. They are never used to train AI models.
Sensitive data protected
Personal data is masked before it reaches the model, secrets are flagged, and a request carrying a blocked marker is stopped.
Start small, prove the value, scale

We deliver Secure AI Workspace as a proven setup in your own AWS account. The gateway runs on ECS Fargate or EKS behind an internal load balancer, reachable only from your private network, with Amazon RDS and Secrets Manager behind it and no internet access for the servers.

Discovery workshop
We map how your teams use AI today, which data sources matter and what your compliance needs are. Outcome: an agreed scope and plan.
Proof of concept
The gateway, guardrails and Claude run in your own AWS account. Outcome: a working setup on your own data.
Pilot group
Real users work with real documents while we tune the policies and track usage. Outcome: proven value and real feedback.
Rollout
All groups are onboarded, device policies are deployed and staff are trained. Outcome: secure AI for every team.
Managed operation
As your AWS managed service provider we run it for you: monitoring, updates and audit support, so it stays secure and up to date.
Let's start with a discovery workshop.

One session to map your AI needs and plan the first step.

Let's Talk
What's included
Everything runs in your own AWS account, delivered and operated by us.
Company sign-in, control by group
Employees sign in with Entra ID, Okta, Google or Keycloak. Your identity provider decides who uses which Claude models; nobody holds API keys. Offboarding ends access within an hour.
Guardrails and data masking
Every prompt is checked before it reaches the model: personal data is masked and secrets are flagged. Answers are checked again on the way back.
Audit trail in your account
Logs are masked, encrypted and tamper-proof, kept 30 days in full and one year as an activity record of who used which model and when. Ready for your auditor.
Monitoring and cost control
Usage, tokens and response times flow to Amazon CloudWatch, Grafana, Datadog or any OpenTelemetry tool. Budgets per user and per group keep AI costs predictable.
Compliance by design
Built for GDPR, the EU AI Act (with the staff training Article 4 requires) and DORA: named operators, audit rights, an exit plan and code you own.
Flexible and future-proof
One gateway serves Claude Desktop, Claude Code and the Agent SDK. Switch region or provider, or fail over to a backup, without reinstalling anything on laptops.
Case study

Sedmi Odjel Delivers AI-Powered Legal Research Platform on AWS

Sedmi odjel built a generative AI solution enabling a Croatian law firm to transform legal research through intelligent court decision analysis.

Read case study