A practical seven-step checklist for understanding and managing AI costs through “tokenomics”, which means linking AI spend to use cases, workflows and business outcomes instead of treating tokens as a technical cost. It guides organisations to start with the use case rather than the vendor, and to measure cost per task instead of price per token by comparing it with what the task cost before AI. It also covers tracking cost per team, use case and workflow, and choosing the right model based on the task and how sensitive the data is. The aim is visibility rather than restriction. The checklist ends with three questions to ask before scaling: does it create more value than it costs, do we understand why, and can we still measure it in six months?