About Monitoring and Managing AI Token Usage
Oracle Analytics Cloud AI features use large language models (LLMs) to process natural language requests, generate responses, analyze data, and support AI-powered workflows. These AI operations consume tokens, which represent units of AI processing. Some AI operations consume significantly more tokens than others. Complex prompts, large datasets, document analysis, and multi-step reasoning workflows can rapidly increase token consumption.
Because AI services incur operational costs, Oracle Analytics Cloud provides token usage monitoring and management capabilities to help administrators track AI consumption, configure usage limits, and manage overage behavior.
Oracle Analytics Cloud provides the Token Usage Settings dialog where administrators can monitor current AI token consumption, review their remaining token allowance, configure usage alerts, control overage spending, and manage how Oracle Analytics Cloud responds when its token limits are reached.
Your organization receives a monthly token allowance. This token allowance automatically renews each month, is shared across supported AI features, and is consumed as users interact with AI in Oracle Analytics Cloud. The number of tokens depends on the compute size of your Oracle Analytics Cloud. See Token Limits for AI Services in Administering Oracle Analytics Cloud on Oracle Cloud Infrastructure (Gen 2).
If your organization consumes its monthly token allowance, the AI-powered features may stop processing requests, users may receive messages indicating that token limits have been reached, and users may need to contact an administrator for assistance.
Note:
To avoid service interruptions, administrators can configure token usage alerts, allow token overages, restrict high-consumption AI features, adjust AI feature access for selected users, or switch AI to use the legacy model, which doesn't charge tokens for generative AI features.
Administrators can configure alerts for administrators and users to receive email notifications when token consumption reaches a specified threshold. Alerts help organizations proactively monitor usage and manage AI consumption before token are depleted. See Configure AI Token Usage Threshold Email Alerts and Configure Token Terms and Billing Settings.
If the administrator enables token overages, Oracle Analytics Cloud continues processing AI requests after the included allowance is exhausted. Additional usage is billed using OCI Generative AI service pricing. See OCI Price List.
Prerequisites
Before you manage token usage settings, ensure that these prerequisites are met.
- You have the BI Service Administrator application role or belong to an application role with the Manage LLMs permission to manage token usage for your organization.
- Your OCI administrator set up AI token billing for additional token usage. See Enable Overage Billing for AI Tokens.