The seven screens

Tokenomics has seven screens. Each one answers one question.

Overview#

Answers: who answers for AI spend, what it returns, and what it costs this month.

• The person accountable for AI economics.
• How much spend sits in a named workload with an owner.
• Value booked per AI dollar.
• Spend against the monthly budget, with a projection and its range.

AI Value#

Answers: what each workload returns.

• Outcome metrics give a cost per outcome: tickets resolved, pull requests merged, units sold.
• A metric can count itself from a Jira or GitHub query, month by month, with the query kept beside the counts. The current month stays partial until it ends.
• Value claims follow the Tokenomics Foundation's chain from an observed change to a dollar result. Only finance, admin and super admin can book one, once, on evidence they name.
• Total cost sets token spend beside the other costs a workload carries: platform and people, licenses, vector databases, evaluation, observability and GPUs.
• Customer-facing workloads show their margin after AI.

Efficiency & Routing#

Answers: how well the tokens are used, and what to change.

• Cache hit rate, and whether caching pays.
• Batch and tier use.
• The model mix, and the spend that runs behind a router or gateway.
• Why the bill moved: volume, mix or rate.
• What a workload would cost on an open-weight model.
• For models you serve yourself: cost per token beside the API price, with utilization, energy and carbon.

Actions turn what the data shows into owned tasks, such as moving work to batch, enabling caching, testing a smaller model or reclaiming idle seats. Each task has an estimated monthly saving and can become a Jira ticket. When a task is marked done, the saving is measured afterwards on your bill, at the same token volume.

Forecast & Planning#

Answers: what the next three, six or twelve months will cost.

• A forecast with a range from backtests, against the budget and by workload.
• The drivers of each workload's cost.
• Scenarios priced against the budget.
• A record of how each past month's forecast did.

Data sources#

Answers: where the numbers come from.

• Where the rows come from, and what each feed can show.
• Which of the six channels of AI spend your data covers.
• Who owns each spend source.

See Data sources and access.

Cost Allocation#

Answers: who spent it.

• Showback by team, workload, person, provider, channel, model, project, key, source, tag, region or environment.
• Workloads are named here, each with an accountable owner, a stage and a sunset date.
• Allocation rules assign new spend to workloads.

Budgets & Governance#

Answers: whether spend is on budget, and what a budget cap cannot catch.

• Budgets in dollars or tokens, with a status based on the projected month end.
• Anomalies grouped into incidents, each naming the source, model or tag that drove it.
• Six policies: models outside an approved list, experiments past their sunset date, unowned spend, new keys spending, runaway loops, and Frontier model share over a cap.
• Developer AI: Claude Code, Copilot and Cursor usage against their seats, by team.
• Alerts by email, to a Slack channel per budget, or to Jira.
• A weekly digest, on the day you choose: the week's spend and change, the month's projection against the budget, the biggest movers, open incidents, unowned spend and one recommended action.

Next#

• What is Finitizer Tokenomics
• Data sources and access