Executive Simulator · FinOps of AI
Operation Cognitive Capital
An interactive simulator that teaches by discovery the real economy of Artificial Intelligence in companies: Why the most expensive AI no is the one that charges most for use — and how to measure what She really delivers.
You don't need to know anything about AI in the first place. Each screen has purpose, step by step and a glossary in simple language.
The central idea
When a company adopts AI, it sees only the tip of the iceberg: the usage invoice (the API). But the real cost is dominated by what is submerged — the human hours spends reviewing and correcting what the AI has produced, infrastructure and governance.
In large business software pipelines, the API is usually less than 10% of total cost. Who optimizes only the price of the token is discussing the wrong part of the problem. The simulator It leads you to rediscover that -- and invent the metric that corrects the error.
What you come out knowing
- 📊 Read the total cost (TCO) of an AI operation — not just the API invoice.
- 🧭 Use UI/$ (Useful Intelligence per Dollar) to compare architectural decisions.
- 🧠 See human cost nonlinear: the fatigue of checking wrong AI (V core).
- ⚖️ Defend Choices with evidence — including in regulated sectors (Governed UI/$).
The journey in 6 chapters
A guided investigation, from mystery to defense before the Council.
- 1 · The Mystery
Where did R$ 37,000 of the AI invoice go?
- 2 · The Trap
Because the AI "barata" is expensive.
- 3 · The Invention
You deduce the efficiency metric yourself.
- 4 · Formalization
The official formulas behind intuition.
- 5 · The Laboratory
Operating the company for four quarters under uncertainty.
- 6 · The Council
Defend every decision with evidence.
Rigor, not achism
🗂
11 evidence
Each concept comes with source, epistemological level and declared limitations.
📐
3 formal models
Cost-of-Pass, UI/$ and Governmentd UI/$ — the math behind the decisions.
🔗
Real references
Links to the fonts (RouteLLM, prompt caching, Context Rot and more) in the glossary.
Honest note: UI/$, V core and such are FinOps operational conventions in consolidation — directional compasses, not ISO/IEEE standards. Simulator numbers are teaching parameters, not absolute measurements.
Who's it for?
Engineers
that implement AI and want cost criterion.
Architects / Tech Leads
who need defensible standards.
Managers / Executives
who decide to budget without entering the code.
Students
They want strictness and verifiable sources.
How it works
🧭
There's always a goal
A banner on top tells you what to do. Click "How to do?" for step by step.
📖
Everything is explained
Help and Glossary translate each term into simple language.
↺
No fear of error
Back or Restart Anytime. It's a lab.
Your mission begins now
You've just been hired as the AI Architect and FinOps of Nexus Tech Corp, a company in crisis for uncontrolled use of AI. You have 24 hours to explain the break — or AI projects are cancelled.
Chapter 1
The Mystery of the R $ 37 thousand Missing
THE CFO Helena Vance: "The API invoice was R$3,000. But the financial points R $ 40,000 in AI costs. R$ 37,000 is gone. Explain."
| Category | Value | % |
|---|---|---|
| API Cost | R$3,000 | 7,5% |
| Time Dev (Review + Verification Load) | R$ 28,800 | 72% |
| Infra / Vector DB / Observability | R$ 4,800 | 12% |
| Governance Overhead / Compliance | R$ 3,400 | 8,5% |
Chapter 2
The trap of the AI "Barata"
THE CTO Dr. Aris Thorne wants to migrate everything to the Model-Lite (90% cheaper for token). THE CRO Marcus Brody. He hesitates, "Legal risk is no joke."
Chapter 3
Invention of Metrics
The board requires a unique indicator of real efficiency. Mount the fraction that captures "useful work by real invested".
Chapter 4
Math formalization
Chapter 5
The lab "What If?"
Take over from Q1 to Q4. Configure the architecture to maximize UI/$ and survive market events. Your decisions today affect the next quarter.
TCO Cost Flow
Historical evolution
Chapter 6
The Council meeting
The counselors consulted the Digital Twin metrics. Defend each decision With an argument and evidence from Mural.