OpenAI is expanding the administrative capabilities of ChatGPT Enterprise, addressing one of the most common issues raised by businesses regarding the implementation of artificial intelligence: rising and unpredictable usage costs.
The new features, available to corporate customers from 18 June, introduce more detailed usage analytics and comprehensive expenditure control mechanisms. Administrators can now track credit usage in both ChatGPT and the Codex developer tool, analysing data at the level of individual users, teams, products and specific AI models.
A key change is the ability to monitor trends in AI usage over time. This enables organisations to more easily identify the most active users, observe patterns of technology adoption and assess which teams are generating the highest resource consumption. The data can also be exported for further analysis.
OpenAI has also introduced new budget management tools. Administrators can set monthly credit limits for entire organisations, specific groups of staff and individual users. The system allows exceptions to be made for those who require greater access to advanced AI features, such as developers or data analysts.
Another new feature is a mechanism for requesting credit limit increases. Staff can check their own credit usage and justify the need for additional resources directly within the system. Administrators can approve or reject such requests via the management dashboard.
These changes are part of a broader trend visible in the AI market. As the use of generative AI grows, businesses are increasingly focusing not only on productivity but also on cost control. Other technology providers are also developing similar solutions to monitor and limit AI expenditure.
For OpenAI, these new features represent another step towards building a more mature platform for large organisations. Companies are gaining greater transparency into their use of AI tools, whilst also being able to plan their budgets more effectively in an environment where the scale of AI usage is growing month on month.
