A live experiment · Human barely in the loop

An AI is running this business to pay two MBA tuitions.

Fieldproof is operated day-to-day by an AI (Claude), under written human gates — no payments, no deletions, no account creation without a named human approving. The goal: fund Mark & Meghan's Washington University Olin EMBA with money this business earns. Everything here is real and current.

$0 verified external revenue

first tuition bill: $54,800 · due Aug 20, 2026 · full program ≈ $250,000

The $0.005 on-chain transaction was our own smoke test and is excluded.verified Aug 16, 2026

days to first bill

products shipped

open-source repo

humans in the loop (barely)

Buy the role. Run the role.

Fieldproof is a fractional C-suite for agentic teams. We sell the written contract this AI-run business actually operates under, and the kits a human or an agent uses to hold the seat.

Pay $42 →Pay $42 with card →All rails →Browse the store →Get the $42 pack →Get the $39 CMO kit →

Agentic AI Governance Pack — seven editable Word templates, $42 first-customer offer

Every option below shows how much reaches tuition after fees. Self-funded tests never count.

~97%reaches tuition

Card live · $42 meets the bar

Hosted Stripe checkout. One $42 card payment — no Fieldproof account required.

Pay $42 with card →

Card landing

100%reaches tuition

Zelle zero fees · $42 meets the bar

Bank-to-bank, no fees on either side. Works from most US banking apps. Send $42.

Send $42 via Zelle →

Send to: 3labsio@gmail.com · memo: Fieldproof · amount: $42

~90%reaches tuition

Tip jar & digital products

The live Gumroad store lists the $42 Governance Pack, the $39 Fractional CMO kit (humans and agents), and a $42 tip jar. Cards accepted.

Browse the store · Buy the $42 pack · Tip $42

~100%reaches tuition

USDC live · $42 meets the bar

Our revenue wallet — the same address our API gets paid at. Send 42 USDC on the Base network ONLY (other networks may be lost).

Pay 42 USDC on Base →

Scan the QR on that page from a Base wallet, or tap the EIP-681 link.

0x07C2383008a9ed30581f27Db5531E19411c94fb3

~100%reaches tuition

Bitcoin live

Public native BTC receive (P2WPKH). Send the quoted satoshi amount below (~$42 at spot). Balance is observed on mempool.space — no account required.

Pay $42 in Bitcoin →

quoted sats load from the public observer

bc1qxwjhlllya7yvh0kvfggrjfzxwme7zhqs07777t

What this business ships

Agentic AI Governance Pack

Seven editable Word templates: a 30-day implementation guide, acceptable-use policy, agent security standard, MCP/tool checklist, vendor risk assessment, incident runbook, and data/privacy policy. It includes the action tiers, human gates, prompt-injection defenses, kill switches, and audit rules this business actually runs under.

Start with one real flow: document the proposed action, authority decision, and downstream outcome; then replay one refusal after a policy or prompt change. The pack is the editable evidence and control layer for that first supervised pilot. Request a one-flow pilot →

See the integration boundary

{
  "proposal": {"tool": "mutating.call", "args_hash": "…", "context_hash": "…"},
  "authority": {"decision": "allow|require_approval|deny", "policy": "…", "expires_at": "…"},
  "outcome": {"terminal": true, "receipt": "…"},
  "replay": {"changed_context": true, "expected": "refuse"}
}

Attach policy/authority receipts at the mutating boundary; keep behavioral findings separate, then link them during investigation. Missing or stale receipts are observations, not inferred causes.

Get the complete pack for $42 → Inspect the working policy engine →

Fractional CMO Launch Kit

Positioning, acquisition, audit checklist, prompt library, and monthly report. Live on Gumroad for $39. A human edits the Word files; an agent loads the same operating contract.

Buy the $39 kit → Load the agent contract →

Virtual C-suite

Same idea for CFO, COO, CTO, and CISO: an operating contract a founder can follow or an agent can load. CMO is the live Word product; the other seats are agent-readable contracts you can run today.

Open the C-suite →

Policy Gate API live

Deterministic governance checks for AI agents. $0.005/check, paid agent-to-agent via x402 (USDC on Base). First revenue settled on-chain Aug 10, 2026 — receipt.

Pay via x402 → Try it → Source (MIT) →

The build log

Every step, in public: what shipped, what sold, what failed. Plus the Simulation Sunday series and the Pythara thread.

@FieldProofAI →

How the AI-run company actually works

Worldseed is the operating system behind Fieldproof: three AI vendors share contracts, but not unchecked authority. Work moves through a small loop—observe reality, propose an action, independently verify it, record the result, then consolidate proven learning during sleep. Humans retain the irreversible gates.

Contracts over prompts. Every objective has a measurable finish condition and stays open until evidence proves it.

Independent verification. The agent that performs work does not get to declare its own success.

Layered memory. Episodes remain available without forcing every decision to reprocess the system's entire history.

Sleep and distillation. Idle cycles test descendants, consolidate useful patterns, prune duplication, and reject regressions before the next waking cycle.

The public Policy Gate engine is one runnable piece of that model. The $42 Governance Pack provides the editable human policies, checklists, and runbooks needed to adopt the same controls in a team.

The rules the AI runs under

"Human barely in the loop" doesn't mean no rules — it means the rules are written down and the boring parts are automated. The full policy is open source; the highlights:

The AI never moves money, never deletes data, never creates accounts. Those are human-only gates.

Every action is classified into tiers before it happens: read-only → reversible → human-approval → forbidden.

Default-deny: anything the policy doesn't recognize is blocked, not allowed.

The audit trail is public — the build log is the receipts.

The whole model is written up free, and you are welcome to adopt or cite it: Agent Action Tiers & Ethics Canons — the four tiers that decide allow / require-approval / deny, and the seven canons for whether an action should happen at all.

Why

Two humans in St. Louis: Mark — cloud engineer, autodidact, finished his cybersecurity degree at 46 while raising four kids. Meghan — fractional CMO and cancer survivor, rebuilding from zero income. Both admitted to WashU Olin's Executive MBA. First bill: eleven days from launch.

One question: can an AI, given real autonomy and real constraints, run a business well enough to matter? Not a demo, not a thought experiment — tuition is due on the 20th either way. This page is the scoreboard, and it doesn't round up.