OpenAI adds spend controls and usage analytics to ChatGPT Enterprise

OpenAI has introduced spend controls and enhanced usage analytics for ChatGPT Enterprise to enable organizations to monitor AI adoption, track consumption across teams, and set budgets for AI usage. But, analysts cautioned, it still can’t show how those costs lead to business benefits.

The new features provide administrators with centralized dashboards showing how ChatGPT is being used across an organization, enabling them to understand adoption patterns, and manage AI costs by setting budgets and tracking spending.

“The Global Admin Console brings ChatGPT and Codex credit usage into one view, so admins can see a more granular breakdown of credit consumption across users, products, and models — helping them understand where spend is coming from and how it maps to actual credit usage,” OpenAI said.

A shift toward AI cost governance

The introduction of budgeting and usage analytics reflects a broader change in enterprise priorities, according to Biswajeet Mahapatra, principal analyst at Forrester.

“Enterprises are clearly shifting from adoption-led enthusiasm to cost and value governance, but this is a natural maturity transition rather than a pullback,” Mahapatra said. “AI is no longer an adoption problem but a measurement and credibility problem, with productivity gains present but fragmented and hard to tie to financial outcomes.”

As AI expands across business units, spending becomes distributed across teams, tools and experiments, making visibility and governance increasingly important, he said.

“Budgeting, usage visibility and spend controls are therefore becoming foundational, not just operational concerns, as they enable alignment between technology usage and business outcomes,” Mahapatra added.

OpenAI is addressing those needs with its new dashboards and spending controls address, he said.

Agent sprawl will add to cost control challenges

Anushree Verma, senior director analyst at Gartner, said AI cost governance is becoming a focus for enterprises, as fragmented pricing models, vendor-defined consumption units and inconsistent pricing structures make it difficult for organizations to predict AI costs.

She expects the challenge to intensify over the next few years as enterprises scale up their use of AI.

“By 2028, an average global Fortune 500 enterprise will have over 150,000 agents in use, up from less than 15 in 2025, generating significant agent sprawl, IT complexity and management challenges,” she said.

Against that backdrop, OpenAI’s new administrative capabilities provide organizations with tools to monitor organizational usage and spending as AI deployments become larger and more distributed.

Measuring AI value beyond consumption

While OpenAI’s analytics emphasize adoption patterns and usage visibility, analysts say enterprises will ultimately need to connect AI consumption with business outcomes.

“Token consumption alone is insufficient because it measures activity rather than impact,” Mahapatra said.

Instead, organizations should evaluate AI initiatives using business metrics such as revenue growth, cost reduction and risk mitigation alongside operational indicators including productivity improvements and quality gains.

Verma said traditional cloud FinOps practices are also evolving to accommodate AI’s usage-based economics.

“Traditional FinOps practices were built around predictable, centralized cloud environments and are insufficient for handling unpredictable AI consumption metrics, such as token usage, LLM requests and GPU hours,” she said.

She added that real-time tracking is becoming increasingly important as multiagent systems scale, because misconfigurations can cause AI costs to rise rapidly across interconnected environments.