AI Transparency

QAWave deploys specialized AI agents for software quality assurance. These agents use large language models (primarily Anthropic Claude) to generate test code, analyze test failures, and propose fixes.

Our agents are classified as limited-risk AI systems under the EU AI Act. They assist human engineers — all agent outputs require human review and approval before being merged into production code.

Every agent output is subject to human review. Generated tests are submitted as pull requests for engineer approval. Proposed fixes are reviewed before merge. Triage suggestions are advisory — engineers decide the response.

Customers maintain full control over which agent actions are automated versus requiring approval.

Portions of this website's content were drafted or refined with AI assistance. All published content is reviewed and approved by a human before publication.

Per EU AI Act Article 50, this disclosure satisfies the transparency requirement for AI-generated content in commercial communications.

QAWave does not train AI models. We use commercially available models (Anthropic Claude) via their API. Customer code and data are never used to train third-party models. Anthropic's zero data retention policy applies to all API interactions.

For questions about our AI practices: ai@qawave.ai. For general privacy inquiries: privacy@qawave.ai.