Finance teams building a Brazil-to-US payroll corridor often frame this as a choice: Pix or SWIFT? The question is understandable but slightly wrong. Pix and SWIFT operate in different parts of the stack — one handles domestic BRL settlement, the other moves USD across international correspondent relationships. In a well-designed corridor, they're complements, not alternatives. Understanding exactly what each rail does, and where it fails, is what lets you build a payroll workflow that doesn't randomly delay on the fifth of every month.
Pix: Instant, Domestic, and Structurally Limited
Pix was launched by the Banco Central do Brasil in November 2020 and operates as a 24/7/365 instant payment infrastructure for BRL transactions within Brazil. Settlement finality is typically under 10 seconds during normal operating conditions. Pix uses chaves (keys) — CPF, CNPJ, phone number, email, or a random alphanumeric string — as the addressing mechanism rather than requiring account and routing number pairs.
For Brazilian companies managing payroll operations, Pix's role is precise: it gets BRL from your conta corrente to wherever it needs to be in Brazil, instantly, at essentially zero incremental cost per transaction. For inbound corridor funding, a Brazilian SaaS company can send BRL via Pix to its FX operator's account using a CNPJ chave, and that leg settles in seconds.
What Pix cannot do is cross the border. Pix is a domestic BRL rail. There is no cross-border Pix — a Brazilian entity cannot send a Pix instruction to a US bank account. The Banco Central has discussed eventual interoperability with foreign instant payment systems (comparable to the ECB's TARGET Instant Payment Settlement exploring linkages), but as of the time of writing, Pix is Brazil-only. Anyone telling you they can "send via Pix" to a US contractor is describing either a domestic Pix leg that feeds into a subsequent cross-border rail, or something that doesn't exist.
SWIFT MT103: The International Layer
SWIFT (Society for Worldwide Interbank Financial Telecommunication) is a messaging network, not a settlement system. The MT103 is the standard single customer credit transfer message used for cross-border wire instructions. When a Brazilian bank sends USD to a US bank via SWIFT, it's sending an MT103 message through SWIFT's network to trigger settlement via correspondent banking relationships — typically through a US correspondent bank that maintains nostro/vostro account relationships with the originating Brazilian institution.
The typical hop structure for a BRL-to-USD SWIFT wire from Brazil to a US bank looks like this:
- Brazilian company's bank originates the MT103
- Brazilian bank's US correspondent bank (often a major New York-clearing institution) receives the message and deducts from the Brazilian bank's nostro account
- If the destination is a smaller US bank not directly correspondent-connected, a further hop through Fedwire or CHIPS may occur
- USD credited to the contractor's US account
Total elapsed time: typically 1-3 business days under normal conditions, longer if there are compliance holds at any correspondent in the chain. Each hop can add a deduction fee — correspondent banks generally charge between $15–$45 per transaction, and those charges are sometimes passed through, sometimes absorbed into the FX spread, and sometimes both.
Two Corridor Designs: The Architecture Tradeoff
When building the BRL→USD payroll corridor, there are two broad architectural patterns in use:
Pattern A: USD Pre-Funded + Pix Last-Mile (Not Cross-Border)
In this design, the operator pre-funds a USD account at a US-side institution. When a payroll run is triggered, the BRL inbound (via Pix from the Brazilian company) arrives at the operator's BR entity. The operator's US entity simultaneously releases USD from its pre-funded balance to the contractor via ACH or wire. The FX leg happens in the operator's treasury, not in the wire chain itself. This structure can achieve same-day USD delivery because the US-side release doesn't wait for SWIFT settlement — it draws on pre-positioned liquidity. The operator carries the FX book between their BRL receipt and USD outflow, managing that exposure with periodic BACEN câmbio settlements.
Pattern B: Pix Inbound → FX → SWIFT → US Bank
In this design, BRL arrives via Pix, triggers an immediate câmbio conversion, and the resulting USD is transmitted via SWIFT MT103 to the contractor's US bank. This is more straightforward to audit end-to-end (one câmbio event, one SWIFT message, clear paper trail), but delivery time depends on SWIFT settlement speed — typically next business day if initiated before the bank's SWIFT cutoff, more likely T+2 if initiated in the afternoon.
| Dimension | Pattern A (Pre-funded) | Pattern B (Pix→SWIFT) |
|---|---|---|
| USD delivery speed | Same-day possible | T+1 to T+2 typical |
| BRL inbound | Pix instant | Pix instant |
| FX exposure | Operator carries book | Locked at conversion time |
| Correspondent fees | Absorbed in operator treasury | May appear as deductions |
| BACEN câmbio filings | 1 per batch (or per run) | 1 per SWIFT MT103 |
| IOF exposure | Single-leg if structured correctly | Single-leg if structured correctly |
Where Each Rail Actually Fails
Pix failures in a payroll context are rare but real. If the receiving chave is inactive, mistyped, or belongs to an account with a CPF/CNPJ mismatch, the Pix returns automatically — usually within seconds to minutes. This is actually an advantage over traditional TED/DOC: the return is fast, and you know immediately. The operational risk is human error on the key, not infrastructure failure.
SWIFT failures are more opaque. An MT103 can be held at a correspondent for compliance screening — OFAC, FinCEN, or local AML review — without any real-time notification to the originating institution. The originating bank learns about a hold when they query the SWIFT GPI tracker, which may not happen until the contractor flags a missing payment. SWIFT GPI (Global Payments Innovation) has improved tracking significantly since its rollout, and most major Brazilian banks now support GPI for cross-border wires, but the transparency improvement doesn't eliminate the hold risk.
We're not suggesting SWIFT is unreliable for payroll — the majority of MT103s settle without incident. The point is that for time-sensitive payroll runs (contractors whose invoices are due on the 5th), a T+1 settlement that gets held at a New York correspondent until T+3 is a non-trivial operational event. It creates contractor relationship friction that's disproportionate to the size of the payment.
What Finance Teams Should Optimize For
The practical takeaway isn't that one rail is better. It's that your corridor design should give you explicit control over the timing risk and explicit visibility into where each leg of the payment is at any given moment.
For the BRL side: Pix is the correct rail. It's instant, it returns quickly on error, and it costs nothing. Any corridor that uses TED for the BRL inbound is adding 2-4 hours of unnecessary wait time on the domestic leg for no benefit.
For the USD side: the question is whether your operator pre-funds (giving you same-day delivery from a pre-positioned balance) or runs live SWIFT on each payroll cycle. For companies paying 10-50 contractors monthly with relatively predictable volumes, pre-funded operators can offer the same-day cutoff that eliminates the SWIFT timing uncertainty entirely.
Ask your provider which architecture they use. The answer determines your real settlement SLA — not the marketing copy on their pricing page.