How we use AI

No usage caps

For the first half of 2026, there are no caps on token usage.

We can't commit to this forever, but right now is a critical window. We're fully migrating to an AI-native way of working, and token costs should not be a blocker.

Go deep. Try new workflows. Break things. Use AI for code, writing, research, debugging, brainstorming. If you're wondering whether you're using it too much, you're probably not using it enough.

Sensitive data

Everyone is free to use whatever AI tools they prefer for day-to-day work. With that said, we encourage towards our default provider for specific use cases.

Specifically, any work involving sensitive data must go through Claude (Anthropic). This includes customer data, financial records, internal metrics, etc. We have organization-level controls that ensure this data is not used for training.