Queryflo Back to Home

Trust Center & AI Ethics

Effective date: January 1, 2026 · Queryflo by Data-Techcon

Responsible AI
Every AI feature is designed to educate, not replace your judgment.
Transparency
AI outputs are clearly labeled and their limitations disclosed.
User First
We never sell data or use your submissions to train models.
Explainability
AI feedback explains its reasoning — not just what, but why.

Our AI Governance Principles

Educational Purpose Only
All AI features are designed to support learning. AI-generated SQL, code reviews, and feedback are explicitly labeled as educational outputs, not production-ready code.
No Black Box Outputs
Every AI response includes explanation — why a query works, what governance issues exist, what the business insight means. We never return a bare answer without context.
Human Judgment Required
We design AI prompts to encourage critical thinking. The AI Judge's 8-step framework, for example, requires users to evaluate the reasoning, not just accept the output.
Data Minimization in AI Calls
We send only what is necessary to the AI API — the challenge schema and user query. We do not send personally identifiable information to Anthropic.
No AI Training on User Data
We do not use your SQL submissions, query history, or AI interactions to train any machine learning model — ours or third parties'.
Bias Awareness
Business scenarios in challenges are reviewed for demographic and representation neutrality. We acknowledge that AI-generated explanations may reflect biases in training data and are committed to identifying and addressing these.

Security Measures

  • All SQL queries execute in a read-only sandbox Postgres role — users cannot write, modify, or delete data.
  • Challenge solution queries are never returned to the client — stripped server-side before any API response.
  • API keys (Anthropic, Supabase service role) are stored as server-only environment variables, never exposed in client code.
  • All data in transit is encrypted via TLS 1.2+.
  • Authentication tokens are handled via Supabase's SSR-safe cookie management.

Reporting Concerns

If you identify an AI output that appears biased, harmful, or factually incorrect, or if you discover a security vulnerability, please contact us at trust@data-techcon.com. We commit to reviewing all reports within 5 business days.