Comparison
Visa protects the budget. SettleProof protects the outcome.
Visa-style pre-purchase guardrails (Stripe Agent Toolkit, Crossmint, Visa Intelligent Commerce) cap how much an agent can spend before the transaction. SettleProof, powered by AAP, closes the loop on what happens after: deterministic settlement when delivery is proven, deterministic refund when it isn't.
Devnet demo output
┌──────────────────────────────────────────────────────────────────┐
│ Scenario 1 — RELEASE (happy path) │
│ tx 4tZ6BfvLAN992q2hvNxnLoirXRXWYeot2FhLLsxFdGmYuh… │
│ merchant: 1 USDC ✓ │ vault: 0 ✓ │ state: fulfilled (100%) │
│ │
│ AAP: ~1.5s (1 tx: Ed25519 ix + release_escrow CPI) │
│ Visa: T+1 to T+3 (1–3 business days for capture/settle) │
├──────────────────────────────────────────────────────────────────┤
│ Scenario 2 — TTL REFUND (dispute auto-resolved) │
│ tx dkNJy9Ugv254jyYt7aRrWV5xT4YWPiZkjeCpjRZ3R2guTAj… │
│ agent: 100 USDC ✓ │ vault: 0 ✓ │ state: refunded │
│ │
│ AAP: < 10s (TTL elapses + 1 claim_refund tx) │
│ Visa: 30–90 DAYS (chargeback lifecycle: dispute, response,│
│ arbitration) │
│ │
│ No chargeback. No arbitration. No 30-day wait. │
├──────────────────────────────────────────────────────────────────┤
│ Scenario 3 — PARTIAL 60/40 (split delivery) │
│ tx 63Z9Reddp7BERHBqryKMyMQB3Xrg3gyUhj3TTxu3J2kuT9gK… │
│ merchant: 0.6 USDC ✓ │ agent: 0.4 USDC ✓ │ state: partial │
│ │
│ AAP: ~1.5s (single CPI tx, on-chain u128 split) │
│ Visa: 60–180 DAYS (partial dispute = full arbitration path)│
│ │
│ On-chain split executed in 1 tx. No mediator. │
└──────────────────────────────────────────────────────────────────┘Side-by-side metrics
| Metric | AAP | Visa / Mastercard |
|---|---|---|
| Settlement (happy path) | ~1.5s | 1–3 business days |
| Refund (no delivery) | TTL + claim_refund tx | 30–90 days |
| Partial dispute | single partial_release tx | 60–180 days |
| Human arbitration | not required for release/refund | pre-arb + arb phases |
| Cost per dispute | Solana tx fee + rent | $15–$50 chargeback fee + product loss |
| Audit trail | cryptographic on-chain | email + PDF + spreadsheet |
| Counterparty risk | merchant cannot move funds without proof | merchant captures funds, user disputes after |
| Customer dispute window | n/a (TTL bound) | 120 days from posting |
| Merchant response window | n/a | 30–45 days |
Why the gap matters
Chargebacks are projected to hit $28.1B in merchant losses by 2026, with global dispute volume rising from 238M (2023) to 337M (2026) — a 42% jump (Ethoca / Chargebacks911 2026 Outlook). Agent-initiated transactions amplify both sides: agents fan out requests at machine scale, and friendly-fraud detection wasn't designed for non-human buyers.
Morgan Stanley Research estimates $190–$385B of U.S. e-commerce will be agent-mediated by 2030. a16z's Sam Broner argues stablecoin thesis frames why programmable settlement rails matter — SettleProof is the attestation layer that makes that practical.
Where SettleProof fits
┌─────────────────────────────────────────────────────────────┐
│ Stripe Agent Toolkit / Crossmint / Visa Intelligent Commerce │
│ ───────────────────────────────── │
│ pre-purchase guardrails │
│ • spending limits, MCC whitelist, human-in-the-loop │
│ │
│ Skyfire │
│ ───────── │
│ agent identity (KYA) + agent-to-API micropayments │
│ │
│ ★ SettleProof / AAP │
│ ───────────────── │
│ post-purchase settlement & dispute resolution │
│ • cryptographic delivery proof (Ed25519 on Solana) │
│ • TTL-bounded refunds, no chargeback │
│ • partial settlement in one tx (no arbitration) │
└─────────────────────────────────────────────────────────────┘Source citations
| Claim | Source |
|---|---|
| $28.1B chargeback losses by 2026 | Ethoca 2023 Outlook (via Chargebacks911) ↗ |
| 238M → 337M global dispute volume (+42%) | Same Ethoca report ↗ |
| Visa chargeback timelines (30–90d / 60–180d) | Visa Compelling Evidence 3.0 ↗ |
| $190–385B agentic e-commerce by 2030 | Morgan Stanley Research ↗ |
| Stablecoins win B2B agent payments | a16z — Sam Broner ↗ |