Dar3 uses a two-track hybrid model. For merchants on enrolled digital channels (SPEI, card, payment link), repayment is architectural — payments route to Dar3's CLABE first, and the platform automatically splits 30% to Dar3 before the merchant ever sees the funds. The merchant cannot skip repayment on enrolled channels because they never receive the gross amount.

For cash economy merchants or non-enrolled channels, repayment relies on domiciliación (auto-debit via Belvo), credit bureau reporting (Buró de Crédito), and wholesaler pressure — the wholesaler's own advance terms depend on their portfolio recovery rate.

The goal is to maximize Track 1 (structural) coverage by selecting merchants with high digital revenue mix and pursuing POS platform partnerships.

Yes — this is a real gap we are solving. On Clip, a merchant can change their settlement CLABE via the app in minutes. The change takes effect after 72 hours. On bank POS terminals (Banorte, BBVA, Getnet/Santander), a CLABE change requires 15 days written notice + acquirer bank written approval — much harder to bypass.

Our defense is layered: (1) STP virtual CLABE monitoring — if no inbound settlement for 48h, an alert fires automatically; (2) contract clause — any CLABE change without Dar3 consent triggers immediate full repayment; (3) 10–15% cash reserve held at closing covers the detection window; (4) Belvo domiciliación fallback — auto-debit fires from the merchant's bank account; (5) RUG pledge — legal priority lien on all POS receivables filed at every closing.

For the pilot, we prioritise merchants on bank POS terminals where bypass requires bank approval, and merchants with >60% digital revenue mix where Track 1 covers most of their volume.

R2 integrates with platforms (Rappi, Uber Eats, inDrive) at the disbursement layer — the platform deducts repayment before paying the merchant. This only works for closed-ecosystem merchants (restaurants on Rappi, drivers on inDrive). R2 faces the same CLABE bypass challenge as Dar3 for general physical POS merchants.

Dar3's roadmap to R2-level enforcement is a commercial partnership with Clip or Conekta — where the POS platform's disbursement engine calls Dar3's API to deduct repayment before paying out. We offer the platform a revenue share (0.3–0.5% of advance volume). This is our primary BD target for months 9–24.

The wholesaler is Dar3's most underrated enforcement lever. The wholesaler was paid by Dar3 on behalf of the merchant. If the merchant defaults, the wholesaler's portfolio recovery rate deteriorates, which directly affects the wholesaler's own future advance terms and referral credits.

This creates a natural alignment: the wholesaler is financially motivated to pressure the merchant to stay current, withhold future orders if needed, and flag early signs of financial distress. This works because the wholesaler has an ongoing commercial relationship with the merchant — a relationship that matters to the merchant beyond just the advance.

This incentive structure is unique to Dar3's GTM model and doesn't exist in direct-lender models like Konfio or Kueski.

The 10% default rate in the conservative scenario is explicitly priced to cover three sources of loss:

Merchant scoring filters for >60% digital revenue mix, which brings real-world expected default below this 10% ceiling. The wholesaler recovery incentive further reduces it. The 10% is the conservative floor, not the expected rate.

Brazil has the strongest enforcement infrastructure in LATAM via the Registradora de Recebíveis system (Lei 12.865/2013 + Resolução BCB 96/2021). All card receivables must be registered with a registradora (CERC, B3, or CIP). Lenders can legally pledge receivables at the registration level.

Effect: The merchant cannot cancel Dar3's claim on registered receivables. It is the legal equivalent of R2's platform-hold model — but available across all card transactions, not just closed ecosystems. This makes Brazil structurally the most enforceable market for Dar3's model, and a key reason for the Brazil market priority in our roadmap.