How to track LLM costs per team and project.
Providers bill AI at the credential. Finance needs the team. How to attribute every model call to a team and a project, before the invoice arrives.
- Topic
- AI cost
- Product
- Novie Control
- Format
- Guide
- Reading time
- 6 min
- Published
- 2 October 2026
- Updated
- 21 October 2026
To track LLM costs per team and project, record the cost of every model call when it runs and attach three tags to it: the team, the project and the workflow that caused it. Providers bill at the credential, so the tags have to be added in the path the call takes. Then reconcile the tagged ledger to each month's invoice, and give every untagged cost a named owner until the untagged share is close to zero.
Why the invoice cannot tell you.
A model provider bills by API key, account or project. It reports tokens and dollars per key and per model. It has no field for the team that ran the workflow, the project the work belongs to, or the business outcome it produced.
One key usually serves several teams. Marketing, engineering and operations share a credential because it was quicker to issue one than three. The invoice then shows a total and hides who caused it, and finance reconstructs the split from logs weeks later, when the month has already repeated itself.
The gap is common and measured. In 2026, 98 percent of FinOps practitioners manage AI spend, up from 31 percent two years earlier, according to the FinOps Foundation's State of FinOps 2026. KPMG's Q2 2026 AI Pulse Survey found 26 percent of large US enterprises had full, real-time visibility into what their AI systems cost to run. In New Zealand, Datacom's 2026 State of AI Index found one organisation in five has a dedicated AI budget with clear cost tracking.
Attribution is where it breaks. In the Tokenomics Foundation's State of Tokenomics, published in September 2026 from 472 responses across 11 industries, visibility and attribution of spend was the second biggest challenge, after proving value. In the 2026 State of AI in FinOps, a survey of 700 engineering leaders and practitioners, only one in five said they could identify the reason within hours if AI spend doubled overnight.
TakeawayThe provider bills by key and model, so the team, project and workflow behind the spend have to be recorded in the path before the call is made.
What to record for every call.
The record is a ledger line, written when the call runs. Ten fields cover it.
- When. The timestamp of the call.
- Team. The team accountable for the spend, taken from the caller's identity rather than typed in by hand.
- Project. The piece of work the spend belongs to, as finance names it.
- Workflow or agent. The process that made the call, so cost can be read per outcome.
- Model and provider. Which model answered, from which provider, in which region.
- Tokens. Input, output and cached tokens, counted separately because they are priced separately.
- Price. The unit prices in force at the moment of the call, so a later price change does not rewrite history.
- Cost. The computed dollar cost of the call.
- Credential. The key or account the call used, which is what the invoice will show.
- Budget. The budget line the call drew down, and who approved that budget.
TakeawayWrite a ledger line when the call runs: who caused it, which model answered, how many tokens at that day's price, what it cost and which budget it drew down.
Where to capture it.
There are three places a call can be counted. The rule for choosing is simple: capture it where the call cannot avoid being seen.
- At the provider. Issue a separate key or project per team. It is quick and it gives a per-team view on the invoice, but it stops at the team: it cannot see the project or the workflow, and keys get shared again within a quarter.
- In the code. Wrap the model library so every call writes a ledger line. It is precise, but it depends on every team adopting the wrapper in every service, including the ones built before it existed.
- In the path. Route every call through one point that adds the tags, records the cost and checks the budget before passing the call on. It is one place to maintain, it covers every team by default, and it is where a cap can stop a call before it is paid for.
Most organisations start at the provider and end in the path. The first gives finance a number this month. The second gives it a ledger it can trust.
TakeawayCapture the cost where the call cannot avoid being seen: start with a key per team, end with one point in the path that tags, records and caps every call.
What your providers' usage data already shows.
Each provider reports usage in its own units. OpenAI's usage and cost reporting groups spend by project, API key and model. Anthropic's Usage and Cost API groups it by API key, workspace and model. On Azure, Microsoft Foundry attributes cost to a project through an automatic tag, in preview. On Amazon Bedrock, application inference profiles with cost allocation tags show spend per application, aggregated by day.
That is enough to see cost by key, project and model inside one provider. The team, the workflow and the business outcome appear only if your keys, projects or tags already say so, and each provider's view stops at its own edge. Use the provider reports to check the ledger, and keep the ledger itself in the path.
TakeawayProvider reports show cost by key, project and model inside one provider. The team and the workflow come from your own tags.
Rolling it out in four weeks.
- Week one: list the spend paths. Every provider account, every key, every card charge for an AI tool and every vendor that has switched on an AI feature inside a product you already license. Ask each team separately; the differences between the lists are the first finding.
- Week two: put the tags in the path. Route model calls through one point that stamps team, project and workflow on each request. Where that is not possible yet, issue one key per team so the invoice splits at least that far.
- Week three: set the untagged rule. Decide that any call without tags is charged to a named person, the owner of the key it used. Untagged spend with an owner falls quickly; untagged spend labelled shared grows.
- Week four: reconcile and publish. Match the tagged ledger to each provider invoice, explain the difference, and send the first per-team and per-project report to finance and to each team lead.
TakeawayList the spend paths, put the tags in the path, give untagged spend a named owner, then reconcile to the invoice and publish.
What finance should receive each month.
- Spend by team, against each team's budget.
- Spend by project, against the business case that approved it.
- The workflows that cost the most, with cost per run and per outcome.
- The untagged remainder and its owner.
- Every budget increase approved in the month, and who approved it.
- The change from last month, with the reason in one sentence.
TakeawayFinance gets spend by team and project against budget, the costliest workflows, the untagged remainder, every approved increase and the change from last month.
How Novie Control does it.
Novie ControlAvailable now
Full cost visibility.
See spend by team, workload and model, before the invoice arrives.
$5,440
spent this cycle.
Example figure.
Sources
- FinOps Foundation, State of FinOps 2026
- KPMG, AI Quarterly Pulse Survey Q2 2026 (24 June 2026)
- Datacom, Datacom releases 2026 State of AI Index research report (August 2026)
- Tokenomics Foundation (LF Projects), State of Tokenomics, Release Candidate 1.0 (23 September 2026)
- Anthropic, Usage and Cost API
- Microsoft, Plan and manage costs for Microsoft Foundry
- AWS, Amazon Bedrock application inference profiles for cost tracking
- Harness, New Harness Report Reveals Enterprise AI Spend Has Outgrown the Systems Built to Track It (29 July 2026)