Finance OS
GTM Finance & Cost Control
Every euro of AI production and ad spend attributed back to the wave and asset that consumed it — with budget caps as a governance gate.
AI production
€418.90
LLM + media gen
Ad spend
€1,390
across campaigns
Creator + infra
€362.30
UGC, storage, MCP
Grand total
€2,171.20
this cycle
Cost by Type
Where the budget is going
Ad Spend€1,390
Creator€350
Video Generation€240
Image Generation€92.50
Llm€86.40
Storage€12.30
MCP Api€0
Cost Entries
Itemised spend attributed to creative waves and assets
| Type | Provider | Wave | Amount | When |
|---|---|---|---|---|
| Llm | Anthropic | CW001 | €86.40 | today |
| Video Generation | Kling | CW002 | €240 | yesterday |
| Image Generation | Higgsfield | CW003 | €92.50 | 2 days ago |
| Ad Spend | Meta | CW004 | €1,390 | 3 days ago |
| MCP Api | Meta API | CW001 | €0 | 4 days ago |
| Storage | Supabase | CW002 | €12.30 | 5 days ago |
| Creator | UGC Creator #4471 | CW003 | €350 | 6 days ago |
Governance
How spend is kept under control
Budget caps gate every launch.
Campaigns above the per-cycle spend cap require human approval before going live. Each AI generation and ad euro is tagged with its creative wave and asset, so cost-per-CW and cost-per-asset attribution is exact — and the CostControlAgent can flag overruns before they compound.
Cost Attribution Notes
Per-CW and per-asset tagging keeps unit economics honest
- AI production cost (LLM + image + video) is rolled up per creative wave to compute a true cost-to-produce before any spend.
- Ad spend is reconciled against attributed revenue to derive blended ROAS at the wave level, not just the platform-reported number.
- Creator licensing and infrastructure (storage, MCP API) are tracked separately so they don't distort creative unit economics.