Trust

Responsible AI

We use AI to build faster, not to cut corners. These are the commitments that govern how Atigan Innovations and its ventures use AI.

Last updated 7 October 2026

  • People stay accountable AI does a large share of our engineering work, but a named person owns every product decision and every release. AI assists; people decide.
  • You know when it's AI Our shop assistant introduces itself as an AI assistant at the start of every chat and answers honestly if asked. When a person from our team takes over, their messages are clearly labelled as coming from a person.
  • Tested and reviewed before it ships Code written with AI help goes through the same checks as any other code: automated type checks and tests, releases that stop when checks fail, the ability to roll back, and regular security reviews.
  • Privacy by design Customer personal data is encrypted in our systems, and we send outside AI services only what a task needs. Customers can delete their own accounts.
  • We follow our AI providers' rules We use AI models, including Anthropic's Claude, only within each provider's usage policy, such as Anthropic's Usage Policy. We review these policies when they change and adjust how we work.
  • No automated high-stakes decisions We do not build AI that makes automated decisions about a person's credit, employment, housing, insurance, healthcare or legal matters. If a future product ever touches these areas, a qualified person will review the output and users will be told AI was involved.
  • Honest numbers Social proof in our shop, such as how many people bought a product, comes from real sales data. We do not publish fabricated reviews or testimonials.
  • Secrets stay secret Passwords and keys are kept out of source code, logs and AI prompts. Access is limited to what each person or system needs.

How we work with AI

AI tools help us plan features, write and test code, review security and document our systems. The founder sets direction, reviews the work and decides what goes live.

Our flagship, Vapor Archive, was built and is maintained this way.

Raise a concern

If you think something we built uses AI in a way that is unsafe, unfair or misleading, tell us at hello@atiganinnovations.com. A person reads every message, and we will tell you what we did about it.