Case Study: Building a Bank Payment Gateway on Salesforce

Engagement Lead
Engagement Lead: Rahul Mehta
Financial Services | Payments & ISO 20022 | Integrations & Security
“Designed for real-time rails and regulatory rigor—secure, observable, and scalable from day one.”

1. Overview & Objectives

A multi-entity UK financial services group needed a modern payment gateway to support B2C and B2B flows across subsidiaries. The solution had to orchestrate initiation → authentication (SCA) → screening → clearing → settlement → reconciliation while complying with PCI-DSS and regional regulation. Strategic objectives: (1) enable real-time payments (Faster Payments/SEPA Instant) with ISO 20022 messaging, (2) centralise fraud & AML screening with configurable decisioning, (3) deliver automated reconciliation back to ledgers and CRM, (4) provide auditability and operational resilience with active monitoring.

Architecture overview: Salesforce + MuleSoft + ISO 20022 rails + Fraud + Reconciliation

2. Delivery Approach

We delivered in three iterative waves to land value safely:

Discovery & Blueprint: Regulatory gap assessment (PCI-DSS, SCA), payment flows, message catalogue (pain.001/pacs.008), non-functional requirements (TPS, latency, RTO/RPO), data lineage & masking.
Foundation Build: Canonical payment model in Salesforce, MuleSoft API-led layers (Experience, Process, System), secrets management, idempotency & retry strategies, secure logging, Salesforce Shield encryption & event monitoring.
Scale & Optimise: Async orchestration with Platform Events, queue back-pressure controls, golden-path 3DS2 journeys, observability (APM, distributed tracing), blue/green release and chaos drills.

Program pillars Pillars: Compliance, Resilience, Speed, Insight, Extensibility. Why it works: Payment steps are decomposed into fault-tolerant services, making compliance changes, scheme extensions, or fraud vendor swaps low risk and fast.

3. Solutions Implemented

1) Financial Services Cloud (FSC): Party & account modelling, mandates, payment preferences, exposure limits, relationship hierarchy across subsidiaries.
2) Payment Orchestration App (Custom): LWC console for payment lifecycle, status tracking, manual review, force-retry & RMA, case auto-creation for exceptions.
3) MuleSoft Integrations: API-led connectivity to clearing partners (FPS/SEPA), card acquirers, KYC/AML & fraud scoring, sanctions lists, and core banking/GL. ISO 20022 transformations with schema validation and idempotent message keys.
4) Security & Trust: Salesforce Shield, field-level encryption, tokenisation for PAN, signed webhooks, SCA/3DS2 step-up, role-based access with platform events for audit.
5) Reconciliation & Ledger: Event-driven postings to sub-ledger, automated receivables matching, dispute/chargeback handling and fee breakdowns; CRM Analytics dashboards for settlement gaps and aging.
6) Operational Excellence: Runbooks, canary health checks, DLQs, auto-heal flows, synthetic transactions, and RTO ≤ 30 min DR plans with data residency controls.
7) Customer & Partner Portals (Experience Cloud): Secure payment initiation, mandate management, downloadable statements, and real-time notifications (email/SMS/Push).

4. Outcomes & Impact

Time-to-authorise: sub-second for real-time rails; < 2.5s end-to-end for 85% of card payments.
Straight-through processing: 92% STP across domestic rails via decisioning and idempotent retry.
Fraud loss rate: reduced by 37% through adaptive rules and step-up SCA.
Ops productivity: 45% fewer manual reconciliations; exception backlog reduced by 60%.
Audit readiness: full lifecycle trace with immutable event logs and SoD controls.
Business agility: new scheme onboarding in weeks via reusable ISO 20022 assets.

“Cloud in India gave us a compliant, real-time gateway with end-to-end visibility. Reconciliation is automated, fraud is down, and we launch new rails without drama.”
— COO, UK Banking Group

Next step for your bank/fintech: Start with a 6-week foundation—map target flows, stand up ISO 20022 assets, integrate one real-time rail, and operationalise observability—then scale to additional schemes and use cases.