Stackpack is the first platform to unify AI token consumption, model attribution, and per-employee usage across Anthropic, OpenAI, and Cursor for finance teams.
Stackpack, the AI that automates, measures, and controls vendors and agents, today announced the launch of AI Token Attribution — the first product built to give finance leaders complete, real-time visibility into how AI is being consumed inside their company, by whom, on which models, and at what cost.
The launch marks a turning point for how CFOs and finance teams account for AI. For the first time, companies can see AI spend as a dynamic operating cost tied to individual employees, teams, models, and outcomes — the same way they think about headcount.
Stackpack gives CFOs the same visibility and accountability for AI spend that they have for headcount, with a clear line to ROI. It’s a new category of financial control for the AI era.”
— Sara Wyman, CEO & Founder at Stackpack
AI Has Changed the P&L. Finance Infrastructure Has Not.
The shift is already underway. Engineering teams are running thousands of Claude and GPT-4 completions per day. Sales teams are using AI agents to draft outreach and summarize calls. Developers are completing code through Codex, GitHub Copilot, Cursor, and Augment. Each of these interactions consumes tokens. Each token costs money. And today, almost none of it is visible to finance.
The typical finance team sees a monthly invoice from OpenAI or Anthropic. They have no idea which team drove the bill, which model was used, what the spend-per-employee looks like, or whether that spend is generating any measurable return.
This is the problem Stackpack’s AI Token Attribution solves.
What Stackpack Now Tracks
Stackpack connects directly to the APIs and billing infrastructure of the tools companies are already using, and surfaces a unified, real-time picture of AI consumption:
OpenAI:
– Total token usage
– Per model token usage
– Per user token usage
OpenAI Codex:
– Token consumption across all plans
– User logins
Anthropic:
– Total token usage
– Per model token usage
– Per user token usage
Anthropic Claude:
– Token consumption across all plans
– User logins
Cursor:
– Total tokens
– Tokens per user/API
For every interaction, Stackpack captures: which model was used, who used it, when, and how much it cost. Finance leaders see this data in real time, not just at the end of the month when the invoice arrives.
A New Line Item in the P&L: The AI Workforce
CFOs have always managed two categories of operational spend: people and software. AI is creating a third category — one that looks like both.
AI agents and AI-assisted employees are now driving real output: writing code, booking meetings, answering customer questions, analyzing data. They are also driving real costs that are not measured in seats or licenses, but in tokens, completions, and compute.
The companies winning the AI transition are the ones that start treating AI consumption with the same discipline they bring to headcount planning: setting budgets, tracking utilization, identifying waste, and measuring return.
Read More on Fintech : Global Fintech Interview with Rob Young, Managing Director – UK at InDebted
Stackpack makes this possible.
“We are at an inflection point in how finance leaders think about the cost of work,” said Sara Wyman, Founder and CEO of Stackpack. “AI agents are doing real work inside companies right now – but almost no finance team can see the cost or the output. Stackpack gives CFOs the same visibility and accountability for AI spend that they have for headcount, with a clear line to ROI. It’s a new category of financial control for the AI era.”
Designed for CFOs, Not Just IT
Unlike API dashboards built for developers, Stackpack’s monitoring agent is designed for finance teams. The interface speaks the language of the P&L — total token usage across all users, token usage per user, token usage per API, and total spend.
Key capabilities include:
Per-employee usage attribution: See exactly how much AI usage each person in the organization is generating, broken down by tool and time period. Identify your highest-consuming users and understand whether their usage maps to output.
Model-level cost tracking: Not all models are created equal — and neither are their costs. Stackpack surfaces which models are being called and how many tokens they use, so finance leaders can ask whether teams are using the right model for the job — or defaulting to the most expensive one by habit.
AI vs. headcount benchmarking: As AI agents take on more work, the ratio of AI spend to human output becomes one of the most important metrics in the business. Stackpack begins to surface that relationship, connecting AI consumption data to the broader workforce picture.
Unified multi-tool view: Finance teams shouldn’t need five dashboards to understand AI spend. Stackpack consolidates OpenAI, Anthropic, and Cursor into a single, continuously updated view, with more platforms coming.
The Beginning of Real AI ROI Management
Today’s launch is the first step in Stackpack’s broader vision: a platform that enables finance leaders to manage the ROI of AI agents, tokens, vendors, and employees together.
The next phase will connect AI consumption data to business outputs. For the first time, CFOs will be able to answer the question that matters most: is our AI investment paying off?
“The CFO of 2026 is being asked to govern a new category of cost,” Wyman said. “It’s variable, usage-driven, and tied directly to product and user behavior. Most existing tools, processes, and mental models weren’t built for it. Stackpack gives finance teams the infrastructure to govern it, and we’re just getting started.”
Catch more Fintech Insights : The AI Shift in Fraud: Why Banks Need a New Playbook
[To share your insights with us, please write to psen@itechseries.com ]
