From Tokens to Firm Value: A Financial Framework for Token-to-Profit Conversion in Generative AI Adoption
Author: Seiryu Ando (Graduate School of Economics, Kyoto University)
Keywords: Generative AI, Firm Value, Token-to-Profit Conversion, AI Utilization Capital, NOPAT, ROIC, EVA, Value Capture, AI Governance
Research Problem
Companies can readily count tokens, model calls, users, and pilot projects. Those usage measures do not, by themselves, show whether generative AI has produced profit or firm value. More activity does not necessarily mean higher productivity, value captured by the firm, or a return that justifies the capital committed.
This study introduces Token-to-Profit Conversion, a financial framework for tracing the full path from technical usage to firm value.
Token-to-Profit Conversion
The framework organizes the conversion process into four stages.
| Stage | Management question |
|---|---|
| AI utilization inputs | How much AI was used, as reflected in tokens, model calls, or implementation projects? |
| Effective task outputs | Did the use produce reliable, reviewable work, better quality, time savings, or decision support? |
| Financial outcomes | Did those outputs affect revenue, cost, or capital efficiency and produce NOPAT that the firm can capture? |
| Firm value | Did the return exceed the cost of AI-related capital and create economic value? |
The framework does not assume that one stage automatically converts into the next. Time saved, for example, creates no captured financial value if the capacity is not redeployed to higher-value work and does not affect revenue or cost.
AI-Related Invested Capital and AI Utilization Capital
The study distinguishes two forms of capital involved in AI adoption.
| Concept | Meaning |
|---|---|
| AI-related invested capital | Financial capital committed to models, data, integration, security, operations, and governance |
| AI utilization capital | Reusable organizational capability accumulated through human review, feedback, codification of knowledge, workflow redesign, and disciplined learning |
AI utilization capital is not simply the fact that a tool has been deployed. It describes the organization's ability to learn from use and repeatedly turn AI into better outcomes. The distinction allows managers to evaluate financial expenditure separately from the organizational capability that may support longer-term advantage.
Connecting AI to Firm Value
Generative AI creates firm value only when use is converted into NOPAT (net operating profit after tax) that the firm can capture and when the return on AI-related invested capital exceeds its cost of capital.
- ROIC (return on invested capital): how much after-tax operating profit is generated by the capital committed
- EVA (economic value added): whether operating profit exceeds the charge for the capital employed
Making this connection explicit shifts AI measurement from “Is the technology being used?” to “Is it producing value above its capital cost?”
Management Implications
Managers should avoid treating the following pairs as equivalent:
- Usage and value
- Productivity and profit
- Expenditure on systems and organizational capability
- Value created for society and value captured by the firm
- Adoption and the organizational ability to keep learning
Investment decisions and KPI systems therefore need to track more than token cost or user counts. Relevant measures include conversion into effective work, the share of benefits captured in profit, additional capital requirements, and the accumulation of AI utilization capital.
Paper
From Tokens to Firm Value: A Financial Framework for Token-to-Profit Conversion in Generative AI Adoption
Posted on SSRN: July 13, 2026