Instant payments are still a relatively small part of the payments industry compared with credit and debit payment volumes. However, they are fast-growing and quickly gaining global momentum, as shown by The Clearing House RTP network in the United States, the European Union’s TIPS instant payment solution, and other countries’ initiatives.
As instant payments continue to become more mainstream, they will require greater interoperability between government and private enterprise solutions. This should help the adoption of ISO 20022, a global standard for payments messaging that makes electronic data interoperable between organizations. ISO 20022 is expected to enhance payment processes through increased data capabilities such as improved data quality, enhanced straight-through processing (STP) and reconciliation, and increased automation and cost-reduction opportunities. Better data quality and flexibility lead to innovation to develop new products and better serve customers.
But for fintech companies and payment service providers, ISO 20022 will create a flood of data that will put more demands on the data architecture that powers their instant payment systems. As a result, they’ll need to address key technology challenges. Conventional data architectures often can’t combine transactional and historical data for analysis, fraud detection, and risk management in real time, as the data is often siloed or stored in separate systems. Older legacy systems are typically designed to access historical data at periodic batch intervals and can’t meet low latency requirements, resulting in unacceptable response times and outdated or incomplete information. This can lead to customer abandonment, payment denial, fraud, and missed cross-sell opportunities.
Instant payment systems rely on data, analytics, machine learning (ML), and artificial intelligence (AI) technologies to provide not only fast, convenient, and secure payments but also personalized products and services, as well as the engaging experiences that today’s consumers demand. Fintech companies and payment service providers need to ensure their systems can scale to meet demand, eliminate data silos and complexity, lower latency, and be available using geo-distributed applications. And this all needs to be done while still being easy for consumers to use and protecting them from fraud. Perhaps not unsurprisingly, during the past year of COVID-19, digital fraud attempts in the financial services industry have continued to grow, increasing 149% globally in the first four months of 2021, a trend that could put a damper on instant payments growth if fraud prevention is not improved.
To build instant payment solutions that enhance customer experiences, increase revenue, and reduce risk, here are three data strategies that fintech companies and payment service providers need to consider.
Power intelligent instant payment systems with real-time decisioning on transaction data
A new generation of application architecture eliminates the wall between transaction processing and analytics. Gartner refers to this as “Hybrid Transaction/Analytical Processing” (HTAP). An HTAP architecture is best enabled by in-memory computing technology to allow analytical processing on the same in-memory data store to perform transaction processing. By eliminating the latency associated with moving data from operational databases to data warehouses for analytical processing, this architecture enables real-time decisioning on live transaction data. However, advancements in real-time decisioning have been constrained by DRAM memory’s high cost and limited capacity.
To be competitive, fintech companies and payment service providers need to implement a hybrid memory database architecture built to run on flash and SSD storage. Another storage technology to consider is persistent memory, such as Intel® Optane™ DC persistent memory, designed for data-intensive applications. This hybrid architecture can power various data-intensive use cases across the payments industry, as it enables high data throughput with low latency and always accurate, consistent data. Importantly, while instant payments are great for consumers, they are difficult to profitably monetize for financial institutions, so the solution must be able to reduce server footprint to minimize operational costs.
Enhance data ingestion and real-time decisioning
To enhance real-time decision-making for identity resolution, fraud prevention, and customer 360-degree profiles, fintech companies and payment service providers need to combine data from different data sources and data silos, both at the edge and existing systems of record. By integrating ML with edge devices, they can tighten the loop between detecting and countering new fraud patterns as well as predicting customer needs and behaviors to provide real-time personalized offers. To ensure they’re not missing any information, they should also design an API layer that can provide flexibility in the types and amount of data ingested.
Prepare ML and AI systems for extreme scale
Rapid advances in ML/AI play a key role in the fight against financial crime. ML techniques used in simulation models help prepare fintech companies and payment service providers for potential fraud and significantly improve existing financial crime detection systems. However, AI systems are constantly hungry for more data, up to petabytes of data, to become smarter to outwit fraudsters. For extreme-scale situations, companies may choose to synthesize real-time streaming transaction data with other data such as geo-location, behavioral, and mobile data. The more data that can be analyzed in real time, the better developers can incorporate more advanced AI/ML algorithms such as neural nets, deep learning, and explainable AI.
Fintech companies and payment service providers recognize that real-time decisioning and transactional analytics are no longer optional and need to be embraced. To survive and grow, they must continuously invent new use cases to promote customer loyalty and profitability while reducing costs and risk.