Ask five teams what their company spends on AI and you'll get five different answers. Is it the OpenAI and Anthropic invoices? The Bedrock and Vertex AI line items? The Cursor and GitHub Copilot seats?
AI is now built into cloud services, developer tools and SaaS products, so what counts as AI spend depends on how each company chooses to track it.
That's where Finout's new predefined Virtual Tag for AI spend comes in. Every Finout account now comes with a ready-made definition of AI spend across your cloud and AI vendors, kept accurate and up to date by Finout as AI services launch and change. Use it as is, or adapt it to fit how your business tracks AI.
Go from Visibility to Action
This feature grew out of a real customer need; an organization that needed to reallocate AI spend across teams.
With your AI spend tagged, you can:
- Allocate AI cost to the teams, products or customers driving it, so the people making AI decisions see what those decisions cost
- Set budgets for AI spend and catch unexpected spikes before they hit the invoice
- Track AI spend trends over time and across providers to see which ones give you the best value
Create Unit Economics KPIs
Once your AI spend is identified and allocated, the next question is whether it’s paying off. Combine the AI spend tag with your business telemetry, such as API requests, tokens processed, active users or customers, to build unit economics KPIs like:
- Cost per AI request or per million tokens
- AI cost per customer or per active user
- AI cost per feature or product line
Tracked over time, these show whether your AI costs grow efficiently with usage and where to optimize.
Why This Is Difficult Today
AI spend is hard to define, hard to find, and always changing. It shows up in more places than ever, and each one is tracked differently:
- AI vendors. With providers like OpenAI and Anthropic, almost the entire bill is AI.
- Cloud providers. On AWS, GCP and Azure, AI services sit among compute, storage, networking and hundreds of other line items.
- Tools with AI built in. Coding assistants like Cursor and GitHub Copilot blur the line between software spend and AI spend.
And the landscape moves fast. New models and capabilities ship every few weeks, and more teams across the business are adopting AI tools along the way.
To see the full picture, FinOps teams have to decide which services count as AI, build their own tags or rules to group them, and keep that mapping current as AI services launch and change. It's manual, it's easy to get wrong, and a mapping that was right two weeks ago can drift out of date without anyone noticing.
How It Works: A Comprehensive View of AI Spend, Ready Out of the Box
The predefined Virtual Tag for AI spend gives every Finout customer a standard way to track, allocate and understand AI cost from day one.
- Already built. The tag is automatically available in Finout, so there's nothing to set up.
- Broad coverage. It goes beyond LLMs to cover AI services on AWS, Google Cloud and Azure, such as Amazon Comprehend and Kiro, plus Anthropic, OpenAI, Cursor and GitHub Copilot.
- Maintained by Finout. We update the mapping behind the scenes as AI services launch and change, so there's no maintenance on your end.
- Works across the platform. Use it anywhere you'd use a Virtual Tag in Finout: dashboards, views, cost allocation and budgets. Finout's predefined AI views now run on this tag too.
Make It Your Own
Because AI spend means different things to different companies, the predefined tag can serve as a starting point, not just a fixed rule. Duplicate it and adjust what counts as AI to match how your business looks at it.
Need to leave out a single service? Add one filter, instead of rebuilding a long list of rules from scratch.
Get Started
As AI takes a bigger share of cloud spend, knowing what you actually spend on it is the first step to managing it. Finout’s predefined Virtual Tag for AI spend is now generally available in Finout.
Learn more here.
Written By
.svg)
