Account-to-account payments skip the card rails entirely: bank to bank, in seconds, 24/7. PIX and UPI proved the volume; PayShap, FedNow and SEPA Instant are the next wave. The open question is whether A2A finally dents the card duopoly.
A2A is the rail; “instant” is the property; the alias and the confirmation message are what made it consumer-grade.
Account-to-account (A2A) payments move funds directly between two bank accounts without a card network in the path. Instant (or real-time) payments (RTP) are the modern form: cleared and made available to the payee in seconds, 24/7/365, with immediate confirmation to both parties.
This is not new plumbing — bank transfers have existed forever. What changed is the combination of speed, availability, a confirmation message, and a simple address (a phone number or alias instead of an account/branch code). That bundle is what made A2A usable at the point of sale and peer-to-peer, where slow batch EFT never could compete.
The strategic stakes are large: A2A bypasses interchange. Where it reaches scale at the till — as PIX did in Brazil — it directly pressures the card networks’ economics.
Push + alias + request-to-pay + ISO 20022 is the four-part recipe. Miss request-to-pay and you have P2P but not commerce.
The canonical real-time flow is a credit push: the payer instructs their bank to send, the scheme resolves the payee, funds and an irrevocable confirmation arrive in seconds, and settlement between banks happens on the scheme’s clearing and settlement infrastructure.
The payer initiates and authorises; money is pushed. There is no chargeback the way cards have one — finality is a feature and a risk.
A directory maps a human-friendly handle (phone number, VPA, ShapID) to the underlying account so users never type bank details. The directory is critical infrastructure.
The payee sends a structured request; the payer approves in their banking app. This recreates a “bill” or “checkout” experience on push rails and is key to merchant use.
Modern instant schemes run on ISO 20022 (pacs/pain/camt). Rich structured data travels with the payment — remittance info, purpose codes, parties — enabling reconciliation that legacy formats could not.
No banking-hours batch windows. Settlement models vary (deferred net vs real-time gross / prefunded), but the payee experience is instant and irrevocable.
Real-time schemes now exist in most major economies. They differ in ownership (central bank vs bank consortium), settlement model, and — crucially — how hard adoption was won.
| Scheme | Market | Operator | Notes |
|---|---|---|---|
| PIX | Brazil | Central Bank (BCB) | The benchmark. Central-bank-mandated, free for consumers, near-universal adoption within a few years. The proof that A2A can beat cards at the till. |
| UPI | India | NPCI | Largest by volume globally — billions of transactions monthly. VPA-addressed, QR-led, interoperable across apps. |
| FedNow | USA | Federal Reserve | Launched 2023; ~1,500+ participating institutions by 2026 but volume still modest. Coexists with the older private RTP network (TCH). |
| SEPA Instant | Eurozone | EPC / banks | Mandatory across the EU from Oct 2025 — banks must send and receive euro instant credits within 10 seconds, with verification-of-payee checks. |
| PayShap | South Africa | BankservAfrica | Launched 2023. ~136m transactions and R100bn+ processed; limits raised to R50,000. Adoption slower than hoped — pricing and bank UX are the friction. |
The lesson across markets: mandate plus zero/low consumer pricing plus a simple alias drives adoption; leaving it to banks’ commercial incentives slows it. PIX (mandated, free) raced; PayShap (bank-priced, optional) crawled by comparison. SEPA Instant’s 2025 mandate is the EU deliberately copying the PIX playbook.
Speed amplifies everything — including mistakes and fraud. The hard problems of A2A are downstream of the payment, not in it.
Push payments are final. Authorised push payment (APP) fraud — tricking a victim into sending — has no chargeback. The UK’s 2024 mandatory APP-fraud reimbursement rules exist precisely because of this gap.
Without strong payee confirmation, a wrong alias sends money to the wrong account, irrevocably. Verification-of-payee / confirmation-of-payee is now mandatory in the EU and UK for this reason.
A technically perfect rail can still fail to scale if banks price it badly, bury it in the app, or give merchants no clean request-to-pay flow. PayShap’s slow uptake is the cautionary tale.
If the scheme guarantees instant availability but settles net later, a participant default between cycles is a real exposure. Prefunding or RTGS settlement mitigates it — at a liquidity cost.
Instant without rich ISO 20022 remittance data leaves merchants matching payments to invoices manually. The rail can be fast and still operationally painful.
A2A’s pitch to merchants is simple: no interchange, instant settlement, irreversible funds. Against that, cards offer ubiquity, credit, rewards, dispute rights and a checkout consumers trust. The contest is decided market by market.
Where A2A wins: P2P and bill payment almost everywhere, and merchant payments where a regulator forced the issue and the UX is good (Brazil, India). Where cards hold: markets with deep card habits, rewards expectations, and weak A2A request-to-pay flows — much of the US and, for now, in-store retail in South Africa.
For South Africa specifically: PayShap has the rail but not yet the merchant flywheel. The 2026 opening of the National Payment System to non-bank PSPs, plus the SARB’s broader modernisation programme, could change the economics — but A2A at the SA till is still a bet, not a fact.
For peer-to-peer, disbursements, wallet top-ups and recurring bills, A2A is usually the right rail today — cheaper than cards, instant, and increasingly expected. Build request-to-pay and you cover most billing use cases.
A2A at checkout pays off where local adoption is real (PIX/UPI markets) or where you serve high-value, low-margin baskets that interchange punishes. Elsewhere, offer it alongside cards rather than betting the checkout on it.
Treating A2A like cards is the classic error: there is no chargeback, so your fraud, dispute and refund logic must be rebuilt around push finality and reimbursement obligations. Skipping verification-of-payee or APP-fraud controls is both a loss vector and, increasingly, a regulatory breach.
PayShap is the rail to integrate for P2P and payouts now; for in-store merchant acceptance, watch the 2026 NPS opening and PayShap’s merchant proposition before committing. Build to ISO 20022 from day one so you are not re-platforming later.