AI Project Policy

Appearance

AI Experimentation and Production Use Policy ​

TL;DR

Explore AI. Use company-provisioned or approved AI accounts and services for company work. Use appropriate data and approved environments. Ask Dev for help through the official Dev request form when you want guidance or resources. Start handoff when the work needs production systems, sensitive data, or real-world impact.

Summary ​

AI is exciting. We want employees to explore it.

Try new ideas. Build demos. Test workflows. Learn what is possible. Share what you discover.

This policy gives everyone a clear path:

  1. experiment freely with appropriate data and approved tools;
  2. use company-provisioned or approved AI accounts and services for company work;
  3. ask Dev for help when you want guidance, resources, or a better approach through the official Dev request form;
  4. hand the work off when it needs production systems, sensitive data, or real-world impact.

The handoff is the point where a promising idea becomes company work that needs the right specialist owners. The original team keeps the product idea and business context. The specialists bring the standards required for production quality, security, reliability, data handling, and operations.

This is a living document. As our understanding of AI usage within the company evolves, this policy will evolve with it. The goal is to encourage safe innovation, help teams dream big and make useful ideas real, and reduce the blast radius of risky mistakes.

What we want ​

We want employees to:

  • explore AI tools;
  • test ideas early;
  • learn by building;
  • use company AI resources;
  • share useful discoveries;
  • ask Dev for advice on ideas and approaches;
  • write a product specification when an idea is clear but the prototype path is not;
  • ask for help when an idea is ready to grow;
  • turn the best experiments into safe, supported company capabilities.

Who owns this policy area ​

Dev is responsible for the engineering, security, SRE, and DevOps parts of this policy. This includes production access, infrastructure, monitoring, incident response, and operational safety.

Legal, compliance, privacy, HR, procurement, and finance may have extra rules for some use cases. Dev will route requests to those teams when their review is required.

This policy is exactly what its title says: an AI Experimentation and Production Use Policy. It gives guidance for AI product and tool innovation by employees outside Dev control. It does not replace other company policies.

Main rule ​

Use this simple rule:

  • Explore freely with appropriate data in approved environments.
  • When the work needs production systems, sensitive data, or real-world impact, hand it off to the domain specialists.

A successful handoff means the product is held to the same standards as other company systems.

What this is not ​

This policy is not intended to block experimentation, require approval for every prompt, or move product ownership away from the team with the business idea.

It exists to make experimentation safe and to provide a clear path when an idea becomes important enough to need production ownership.

Examples that do not need handoff ​

These examples can usually continue as normal experimentation when they use approved tools and do not include sensitive data:

  • using AI to rewrite an internal announcement;
  • asking AI to explain public documentation;
  • generating mock data;
  • building a local demo against fake APIs;
  • asking AI to summarize a non-sensitive meeting note.

What this policy covers ​

Read these pages in order: