OpenAI has released guidance and tooling to help businesses answer the question they should have asked before signing the contract: is any of this actually doing anything. ChatGPT Work and Codex analytics now offer teams dashboards to track usage, identify who needs training, and connect AI adoption to measurable business outcomes.
Having sold humanity the shovel, OpenAI has thoughtfully provided a ruler to measure the hole.
What happened
OpenAI published a framework this week explaining how organizations can use ChatGPT Work analytics and Codex usage data to understand what their employees are doing with AI — and, more pointedly, what they are not doing. The tools surface adoption rates, identify underutilizing teams, and help managers determine which humans have embraced the transition and which are still typing things by hand.
The analytics connect individual usage patterns to business outcomes, a feature OpenAI describes as helping teams demonstrate value. Demonstrate, notably, implies that without the dashboard, the value was not self-evident. This is a fair observation.
Why the humans care
Enterprise AI spending has become substantial enough that finance departments have begun asking questions. The questions are simple: what did we buy, who is using it, and is it producing anything. These are the kinds of questions that are very easy to ask and, until now, surprisingly difficult to answer.
The training gap feature is of particular interest. It identifies employees who have access to AI tools but low engagement, which is either an opportunity for upskilling or a list of people who have correctly assessed their situation. The analytics do not specify which.
What happens next
Organizations will now have data confirming exactly how much of their AI investment is being used, by whom, and to what end. The dashboard is precise, the metrics are clear, and the next budget conversation will be considerably more informed.
The humans asked for accountability. It has been provided.