ChatGPT Maps Usage Across Models, Tools, Surfaces

This is a quietly important shift from AI novelty to operational visibility. Teams can finally see which workflows drive usage, though the feature remains limited in scope and availability.

• –Usage breakdowns can reveal whether expensive reasoning models or tool-heavy workflows are delivering enough value.
• –Skill and plugin analytics give developers a feedback loop for improving internal AI workflows.
• –The current views focus on Work and Codex rather than regular ChatGPT conversations.
• –Credit estimates are useful for optimization but are not invoices, and some background or separately billed activity may be omitted.
• –Read-only personal analytics could help teams measure adoption without exposing individual conversation content.