The AIW Token Usage Limit feature enables organizations to control and monitor AI usage by defining a configurable token limit at the customer level. This ensures fair usage, predictable consumption, and transparent tracking across all users within a customer account.
Token limits help administrators manage AI-assisted interactions efficiently while ensuring that all users operate within the allowed consumption boundaries.
Note
The token usage policy applies uniformly across all personas within a customer environment.
Customer Level Configuration
Token limits are configured centrally in the backend for each customer. Every user under the same customer account shares the same configured token limit.
Example:
A customer is allocated 10 million (10M) tokens per month and this limit applies uniformly to all users belonging to that customer.
User Level Consumption Tracking
Although the limit is configured at the customer level, token consumption is tracked individually per user. This provides visibility into how tokens are being utilized across different users.
Example:
Customer monthly limit: 10M tokens. User 1 consumes: 2M tokens and User 2 consumes: 5M tokens. Remaining pool continues to be available until the monthly reset.
Note
Limit applies to each user separately.
Example:
A customer is allocated 10M tokens. If User 1 consumes 5M tokens and User 2 consumes 3M tokens, this does not mean the remaining users are limited to the remaining 2M tokens.Each user is entitled to the full 10M token limit. As a result, User 1 has a remaining balance of 5M tokens, User 2 has 7M tokens, and all other users continue to have their own 10M tokens available.
Monthly Token Refresh Cycle
Token usage follows a monthly refresh cycle at the start of month.
At the start of each new cycle:
Token availability is reset to the configured customer limit
Previous month’s consumption does not carry forward
This ensures predictable and controlled AI usage every month.
Note
Token limits are not configurable per user—they are shared across all users within a customer.
Once the monthly token limit is consumed, further AI usage may be restricted until the next refresh cycle.
Exact enforcement behavior (warnings, soft limits, or hard stops) depends on backend configuration.