AI Usage Policy
How Dazvix uses AI-assisted tools in research, design, engineering, QA and operations, and the human review that stays in the loop.
- Updated
- October 6, 2026
Where AI may be used
Dazvix may use AI-assisted tools to support research, drafting, code exploration, test planning, documentation, summarisation, accessibility review, image ideation and operational workflows.
AI output is treated as a draft or assistant layer. Human review remains responsible for final decisions, client-facing deliverables and production changes.
Client and confidential data
We avoid placing confidential client data, private credentials, production secrets or sensitive personal information into AI tools unless the use is explicitly approved and appropriate safeguards are in place.
When AI can help on a client project, we prefer low-risk inputs such as anonymised examples, public documentation, synthetic data or isolated technical snippets.
Engineering standards
AI-assisted code must meet the same standards as human-written code: review, testing, security checks, maintainability and fit with the existing architecture.
We do not accept AI output blindly, and we do not use AI as a substitute for domain judgment, threat modelling, accessibility responsibility or client approval.
Disclosure
If AI use materially affects a deliverable, workflow or risk profile, we will discuss it with the client in plain language. We are comfortable using AI, but not hiding it.