AI Spend Summary

the ai spend summary page is the executive and operational entry point for understanding how ai usage translates into real cost across your organization everything on this page is derived from usage captured at the ai gateway and priced using your configured rates spend overview (mtd, qtd, ytd) at the top of the page you see three headline metrics spend mtd total ai spend for the current month spend qtd total ai spend for the current quarter spend ytd total ai spend for the current year each card includes the absolute spend value a trend indicator showing change versus the previous comparable period these numbers answer a single question quickly how much are we spending on ai right now, and is it accelerating or slowing down? time range and scope all data on this page is scoped to the selected time range, controlled by the date picker in the upper right supported ranges week to date month to date quarter to date year to date last 7, 14, or 30 days custom date range changing the time range recalculates every chart and table on the page there is no cached rollup ai spend, usage, and cost overview this section highlights where usage and cost are concentrated top spending business units shows the business units generating the highest ai spend in the selected timeframe business units are defined by your attribution model spend reflects real usage priced with your configured rates trend indicators show relative change over time this answers which teams or applications are driving ai cost? top models by usage (tokens) ranks models by total token usage, regardless of cost key point high usage does not necessarily mean high spend lower cost models can dominate usage while contributing less to total cost this is useful for understanding adoption patterns identifying heavily used models spotting opportunities to shift workloads top models by spend ranks models by total cost, not usage this highlights expensive models with lower usage cost concentration risks models that dominate budget impact used together, top usage by model and top spend by model expose pricing inefficiencies and model selection issues spend by ai provider this chart breaks down total spend by ai provider, such as anthropic openai amazon bedrock azure ai others as configured this view is critical for provider level budgeting vendor negotiations detecting provider concentration risk spend here reflects your actual configured rates, not public list pricing spend per request by model this view shows the average cost per request for each model the calculation incorporates average input tokens average output tokens model pricing this answers which models are expensive every time they are called, not just in aggregate? this is one of the fastest ways to identify inefficient prompt designs overuse of premium models opportunities to downgrade models without losing capability usage and spend by business unit (detailed table) at the bottom of the page is a detailed breakdown by business unit for each business unit you can see input tokens output tokens total tokens total spend this table is the foundation for internal chargebacks showback reporting accountability across teams all values respect the selected time range and pricing configuration how this page is meant to be used use the ai spend summary to get an immediate snapshot of ai cost health identify which teams, models, and providers matter most understand whether usage or pricing is driving spend decide where deeper investigation is needed this page tells you where to look next, not everything at once drill into analytics, observability, or chargebacks once you see a signal worth investigating common misinterpretations to avoid high token usage does not imply high cost low total spend does not mean low per request cost provider spend shifts often reflect model routing changes, not usage spikes flat spend can hide growing inefficiency if request size is increasing