# EU AI Act

EU Artificial Intelligence Regulation

## Our Position Under the EU AI Act

Under the EU AI Act, Sistava is a limited-risk, general-purpose AI assistant. That means our obligations are about transparency, making it clear you are working with AI, what it can and cannot do, and keeping you in control, not the heavier requirements placed on high-risk systems like credit scoring or biometric identification.

We proactively align with the Act's principles because we believe responsible AI deployment is a competitive advantage, not a burden.

## Transparency and Disclosure

Users always know they are interacting with an AI system. Our platform clearly identifies AI-generated content and actions. We never disguise AI employees as human workers.

We document the capabilities and limitations of our AI agents. Every employee profile shows which AI model powers it, what tools it has access to, and what guardrails are in place.

## Human Oversight

Our approval gateway ensures that sensitive actions require human authorization before execution. Guardrails prevent AI agents from taking actions outside their defined scope.

Team leaders and company administrators have full visibility into AI decision-making through activity feeds, execution inspectors, and cost tracking dashboards. No AI action happens in the dark.

## Independent Self-Assessment

Sistava completed the EU AI Act Compliance Checker provided by the Future of Life Institute (an independent non-profit). The assessment confirms Sistava operates as an EU-market AI Provider, with no prohibited practices, no high-risk use cases under Annex I or Annex III, and transparency obligations under Article 50.

Assessment date: 26 May 2026. The shareable result remains available for public verification.

## What this means for customers

- AI interactions are always clearly identified as AI
- Human approval required for sensitive actions
- Full audit trail of all AI decisions and actions
- Guardrails prevent AI from exceeding defined scope
- Transparent documentation of AI capabilities and limitations
- Independent self-assessment publicly verifiable