This is an excerpt from the Future of Payments 2025 report.
Technology has never evolved as rapidly as it does today. Finance and banking are ripe for innovation. AI, tokenization, stablecoins, CBDC, and hyper-personalization are all factors in the discussion when it comes to making transactions more secure, streamlining operations, reducing costs, and improving customer experience.
Let’s examine these trends in more detail.
AI in banking and payments
Payment service providers (PSPs) have long used rules-based AI, such as machine learning and robotic process automation. Now, new forms of AI, centered around large-scale language models (LLMs) or generative AI, are enabling new use cases and providing innovative differentiation opportunities for institutions.
One of the key use cases is in AI’s fraud detection and mitigation capabilities. Many organizations have begun implementing AI-based AML and fraud solutions. This helps improve detection rates by reducing false negatives or positives and reducing the manual effort required to mitigate modern threats.
Additionally, the increased availability of generative AI could improve know-your-customer (KYC) processes by allowing financial institutions to reduce the time needed to investigate potentially suspicious activity.
Amelia Ruiz Heras, Head of Global Solutions Consulting, Payments at Finastra, emphasized: “This technology provides automated sourcing and verification of customer information, reports anomalies to account managers, and evaluates customer-specific market data and media stories to ensure KYC. Standards continue to be met. With other features such as voice payment authentication, improved counterparty screening, and cross-checking more data records in a shorter period of time, this technology has several features for robust payment security. It has some promising uses.”
Synthetic data is another opportunity unlocked by AI. By 2025, data will be one of the most valuable assets in any organization, and agencies will not only be faced with vast amounts of data that they need to access and protect, but also to train their anti-fraud models. It also requires complex and comprehensive data. . To protect sensitive customer data, synthetic data is one of the most effective ways to address data issues.
“Synthetic data generated by AI is also being explored to provide financial institutions with a more comprehensive and realistic business and market outlook, thereby improving the accuracy of cash flow forecasting. It can also be used to detect types of fraud, allowing financial institutions to stay ahead of fraudsters as techniques continue to evolve,” added Luis Heras.
When looking at the back office, AI has the potential to improve operations, strengthen security, and increase productivity, helping to significantly reduce costs for organizations. Meanwhile, in the front office, AI can drive revenue by opening up new revenue streams and providing access to previously untapped potential.
Looking to the front office, hyper-personalization is one of the key opportunities unlocked by AI. Hyper-personalization enables organizations to meet increasingly complex customer expectations by delivering highly customized products, streamlining customer service journeys, and even improving security.
Sourav Agarwal, Head of Global Payments at Accenture, pointed to the importance of ISO 20022 in these efforts, saying, “Its enhanced data format allows for richer data processing, such as structured remittance data and purpose codes. Enables the exchange of information, which increases straight-through processing (STP) speeds, reduces errors, and provides greater insight for businesses, reducing fraud and global payments. The system becomes more efficient and allows businesses to create a better experience for the end consumer.”
In conjunction with the increase in data that will be made available with ISO 20022, organizations will have access to a wealth of information about their customers across a wide range of financial situations. Combining available data with AI and behavioral analytics can help fight fraud by knowing what normal and abnormal customer transactions look like.
However, there are several hurdles to overcome to achieve this, says Kevin Flood, FIS Payments Ecosystem Strategy Director for Corporate and International Banking at FIS Global. It is still not fully regulated, so it remains at a high level. Risk appetite remains somewhat cautious as the impact of recent AI regulations will take time to unravel and implement.
Nevertheless, the use cases for AI continue to expand, and the relative certainty of regulation will allow us to push the envelope. The use of AI and GenAI in hyper-personalized products should be a responsibility, along with its use to combat the unstoppable rise in fraud. ”
Turning to fraud risks, Mr. Agarwal points to global initiatives such as beneficiary verification in the UK, multi-stage fraud detection in Sweden, the launch of transaction monitoring in the Netherlands, and Malaysia’s recent national fraud portal. He pointed out that many different approaches have been taken. Still, he said, “the ideal approach would involve all stakeholders in the industry (payments market infrastructure, banks, telcos, social platforms, law enforcement, regulators) and engage in fraud prevention, fraud reporting, collective action and “Focus on the entire value chain of fraud.” Recoveries and refunds to customers. Cross-industry collaboration through shared fraud databases and real-time information exchange allows organizations to stay ahead of new techniques introduced by fraudsters. ”
Given that AI technology is in its infancy, Ruiz Heras outlined four steps institutions should take before deciding whether to implement the solution.
Decide how to measure the benefits that AI can bring, such as improving operational efficiency and straight-through processing (STP) while reducing TCO and fraud losses. Combine these metrics with softer benefits like improving the client experience, having an easier way to innovate, and building your brand with improved AI-generated materials. Work with technology partners to explore alternative ways to achieve your goals. We assess both the nominal investment associated with each path and the level of disruption and risk each path imposes. Consider the risks of rapidly evolving technology, keeping in mind that some of AI’s most profitable applications are probably yet to be discovered. Ask your technology partner to continually evaluate new use cases. Be sure to understand the AI usage policies and processes your partner has established within their organization, as well as any regulatory requirements they must comply with, such as EU AI laws.
Challenges and benefits of CBDC, tokenized deposits and stablecoin adoption
Tokenized money is on the horizon in many parts of the world. As of September 2024, 134 countries and monetary unions representing 98% of global GDP are considering CBDCs. To date, three CBDC schemes have been launched (in the Bahamas, Jamaica and Nigeria) and 44 schemes are in pilot stages, including most of Europe and the Asia-Pacific region. The North American CBDC scheme is still under development.
While development continues to advance, Annalisa Ludwinski, Head of Correspondent Network Management at Investec, describes the challenges of the situation: Regulators and central banks cannot respond. Connecting to multiple private loop networks already has costs and challenges, and there is little or no interoperability between them. This problem will only get bigger. ”
Naveen Marella, co-head of Onyx at JPMorgan, added: “Market fragmentation is inevitable as different parties develop and test different innovative solutions.” Alternating divergence and convergence are a natural part of innovation. Better solutions will likely yield better results, but the key is to drive convergence quickly and minimize the period of fragmentation. ”
One concern is the programmability of tokenized assets. Petia Niederlaender, Director of Payments, Risk Management and Financial Literacy at Oesterreichische Nationalbank, said programmability raises privacy concerns, can lead to new operational and cybersecurity risks, and poses new challenges for regulators. He stressed that this could not only lead to a negative impact on economic stability, but also have a negative impact on economic stability.
“In this context, it is important to distinguish between ‘programmable money’ and ‘programmable payments or smart payments.’ ‘Programmable money’ actually has rules that determine how the money is spent. “In contrast, ‘programmable or smart payments’ are conditional payments that are executed automatically when certain predefined conditions are met,” Niederlaender explains.
But this is also where innovation will play a key role and where a digital euro could become a key driver of future progress, she added. “As the primary payments infrastructure, the digital euro will be a fundamental railing for fintech companies and startups to innovate new use cases, value-added services, convenient consumer and business solutions, and While the core features of a digital euro would ensure broad public access and inclusion of vulnerable groups, private sector market participants could make conditional payments via APIs. We can leverage this by offering enhanced services and value-added solutions that include:
Increased efficiency is not the only benefit that programmability can provide. Additionally, tokenization increases access to financial services and democratizes the sector, especially for unbanked and unbanked groups.
Finally, Ludwinksi added the positive effects on global trade related to programmability. “There are certainly benefits if we consider the development of stablecoins to help facilitate global trade. Being able to link to smart contracts that automatically make payments could provide significant operational efficiency savings as well as a better way to manage risk and exposure.”
