The Ten-Second Build: an EMI's instant-sending programme

Executive summary

An electronic money institution (EMI) engaged the hub to build an instant-sending programme ahead of the 9 April 2027 SEPA Instant deadline. The programme delivered 24/7 payment operations, a ten-second processing window, and a real-time liquidity bridge. Go-live was Q1 2027, ahead of the regulatory deadline.

Client details are anonymized per the Ch. 10 labeling regime. No reasonable reader can reverse-engineer the client from the information published here.

Client context

The client is a euro-area EMI with a payment services portfolio spanning remittance, merchant acquiring, and B2B supplier payments. The EMI's existing payment infrastructure was batch-oriented: payments were queued through the business day and settled in end-of-day batches via STEP2. The EMI had no instant payment capability.

The EMI's regulatory perimeter placed it squarely in the 9 April 2027 sending obligation. Unlike banks, the EMI did not have an existing TIPS or RT1 connection. Everything had to be built from scratch.

Business challenge

The 9 April 2027 deadline requires three capabilities that the EMI did not have:

1. 24/7 payment processing. SCT Inst operates 24 hours a day, 365 days a year. The EMI's operations team worked business hours, Monday to Friday. No weekend coverage. No night coverage. No holiday coverage.

2. Ten-second processing window. Each payment must be processed, fraud-screened, sanction-screened, liquidity-checked, and submitted to the clearing infrastructure within 10 seconds. The EMI's batch processing pipeline took 90 seconds per payment for fraud and sanction screening alone.

3. Real-time liquidity management. SCT Inst settles individually, not in batch. The EMI's treasury function funded the STEP2 account once per day. Real-time liquidity monitoring and top-up did not exist.

Approach

The programme ran in three parallel workstreams over 14 months:

Workstream 1: Clearing infrastructure integration

Connected the EMI to TIPS as the primary clearing infrastructure, with RT1 as backup. Built a payment router that could fail over between TIPS and RT1 without dropping in-flight payments. The router was tested with 10,000 payments in a production-equivalent environment before go-live.

Workstream 2: Processing pipeline rebuild

Replaced the batch fraud screening and sanction screening pipeline with a streaming pipeline that processes each payment in under 3 seconds. The pipeline uses a rules engine for first-pass screening and a machine learning model for second-pass review. Payments that clear both passes are submitted to the clearing infrastructure. Payments that flag for review are routed to a human analyst queue.

The human analyst queue is staffed 24/7. The EMI outsourced night and weekend coverage to a managed operations provider. The managed operations provider follows a runbook written by the hub team. The runbook covers 47 exception scenarios.

Workstream 3: Liquidity bridge

Built a real-time liquidity monitoring and top-up system. The system monitors the TIPS account balance every 5 seconds. When the balance drops below a configured threshold, the system automatically triggers a top-up from the EMI's prefunded euro account at a tier-1 bank. The top-up is executed via an API call, not a manual transfer.

The liquidity bridge was the highest-risk component. A failure in the bridge would mean the EMI could not send payments. The bridge was built with redundant bank connections: if the primary bank API is unavailable, the system falls back to a secondary bank. If both are unavailable, the system alerts the treasury team and pauses outbound payment submission.

Solution

The instant-sending build went live in Q1 2027, six weeks ahead of the April deadline. The first week of production processing showed:

Metric	Value	Notes
Processing window (median)	4.2 seconds	Well within the 10-second requirement
Processing window (P99)	8.7 seconds	Within the 10-second requirement
Peak throughput	340 TPS	Exceeded the 200 TPS design target
Fraud screening pass rate	97.3%	2.7% routed to human review
Liquidity bridge activations	12 in week 1	All successful, no manual intervention
Clearing infrastructure failover	0	TIPS stable throughout

Results

• Go-live: Q1 2027, six weeks ahead of the 9 April deadline
• Processing window: median 4.2 seconds, P99 8.7 seconds
• Peak throughput: 340 transactions per second
• Zero regulatory incidents in the first 90 days
• 24/7 operations maintained through weekends and holidays with no unplanned downtime

Lessons

1. Start the clearing infrastructure integration first. It has the longest lead time. The TIPS onboarding process took 4 months, driven by the ECB's validation and testing requirements.

2. Outsource 24/7 operations if you cannot staff them. The EMI could not hire a 24/7 operations team in 14 months. Outsourcing to a managed operations provider was the pragmatic choice. The runbook is the critical deliverable, not the staffing model.

3. Build the liquidity bridge with redundancy from day one. A single bank connection is a single point of failure. The bridge must have at least two bank connections, and the failover must be automatic.

4. Test with production-equivalent volumes before go-live. The 10,000-payment test in a production-equivalent environment caught three integration defects that would have caused go-live failures. Testing with 100 payments would not have caught them.

---

Anonymized case study. Client permission obtained per the Ch. 10 5-step workflow. Identifying details reviewed with the client. No reasonable reader can reverse-engineer the client from the information published here. Re-verification date: August 2027.