To help you track usage, control costs, and prevent misuse, Userflow applies AI Assistant limits at two levels: assistant-level and global.
Assistant-level limits
Assistant-level limits are configured in the settings for specific AI Assistants within a single environment (e.g., staging OR production). Usage is assessed on a rolling 30-day period.
.png)
What they control: Individual AI Assistant usage.
Scope: Single assistant in one environment (e.g., staging OR production).
When the assistant-level limit is reached:
❌ That specific assistant disappears from your Resource Center for 24 hours.
✅ Other AI Assistants continue working normally.
✅ Test chats and other assistant-related UI functionality remain available.
⚠️ Important: You can still accumulate usage that counts toward billing.
Additional Billing & Auto-Upgrades
Even if your assistant-level limit is reached for a specific assistant and it is removed from your Resource Center, you can still use other AI Assistants and assistant-related UI functions like test chats. These will continue to contribute to your message count and may result in additional billing or auto-upgrades.
Note: While it’s still possible to be auto-upgraded, we’ve built backend protections against auto-upgrades from temporary usage spikes. Consistent usage patterns over the 24-hour period are more likely to trigger an upgrade.
Global limits
Global limits apply to all AI Assistants and assistant-related UI functions (e.g., test chats) in all environments and are managed within the AI Assistant UI. Usage is assessed on a rolling 30-day period and will be displayed at the top of the AI Assistant page, allowing you to monitor your usage and adjust your global limit as needed. If your global limit is reached, all AI Assistants and associated UI functionality that may contribute to message counts will be deactivated for 24 hours. Global limits can be set by account owners..png)
What they control: AI Assistant usage for all assistants and assistant-related UI features across your entire account in all environments.
Scope: Every assistant and assistant-related UI function that contributes to your message count (e.g., test chats, staging environment activity).
When the global limit is reached:
❌ All AI Assistants become unavailable for 24 hours.
❌ Test chats and other assistant-related UI functionality become unavailable for 24 hours.
✅ No further usage accumulates (protecting you from unexpected charges).Additional Billing & Auto-Upgrades
When the global limit is reached, all of your AI Assistants and all related UI functionality that contributes to the message count are deactivated, reducing the possibility of additional billing or auto-upgrades.
Note: While it’s still possible to be auto-upgraded, we’ve built backend protections against auto-upgrades from temporary usage spikes. Consistent usage patterns over the 24-hour period are more likely to trigger an upgrade.
Key Takeaways
Assistant-level limits help you manage individual deployments, but don't prevent overall usage from growing.
Global limits act as a safety net, completely stopping AI Assistant-related usage when reached to prevent unexpected billing or auto-upgrades.
Usage calculations for both assistant-level and global limits are assessed on a rolling 30-day period. If an assistant or related functionality is deactivated when a limit is reached, it will be restored after 24 hours when the 30-day calculation resets.