THE INTELLIGENT BANK: A BLUEPRINT FOR RESILIENT, AI-DRIVEN FINANCE
Jason Cao, CEO of Digital Finance BU, Huawei
INVISIBLE BY DESIGN: THE DISAPPEARING ACT OF PAYMENTS
Savas Manyasli, Chief Technology Officer, Finance Incorporated Limited (FIL)
HOW TO KNOW IF YOU’RE READY FOR PAYMENT ORCHESTRATION
Brady Harris, Chief Executive Officer, IXOPAY
John Trapani, Industry Lead, Financial Services, Appian
THE INVISIBLE REVOLUTION IN FINANCE
Dear Readers,
It’s a pleasure, and a bit surreal, to welcome you to the Summer 2025 issue of Financial IT as your new Editor. My journey here has been far from linear. In my previous role as Editor of Daryo.uz , I had the opportunity to explore the intersection of geopolitics, regulation, and emerging technologies. One highlight was a compelling conversation with Huawei’s Aloysius Cheang on cybersecurity and digital sovereignty. These are issues that increasingly intersect with fintech. That conversation sharpened my sense that we are not just watching finance evolve; we are watching it reinvent itself.
Nowhere is that reinvention more visible than in our cover story, where John Trapani of Appian delivers a powerful, practical, and future-forward vision of AI in payment processing. Titled “Efficient Investigations, Happier Customers,” the article demonstrates how artificial intelligence is no longer an optional upgrade. It is a strategic imperative for any institution aiming to compete in a real-time, crossborder, customer-driven marketplace.
Trapani begins with a jarring statistic: 99.6% of sanctions screening alerts in some financial institutions are false positives. That is not just inefficiency; it creates friction, fatigue, and a real risk to customer satisfaction. With AI-powered automation built into alert investigation platforms, banks and regulators can now streamline case management, shorten resolution times, and reduce the need for rework. At the same time, they maintain audit-readiness and compliance integrity.
But this is not just about speeding up back-end processes. It’s about transforming the customer experience. Through natural language processing (NLP), AI-enabled chatbots and digital assistants are now fielding queries around the clock. Predictive models are helping detect fraud in real time. Agentic AI is supporting customer service reps with instant, data-driven suggestions. And for the first time, financial institutions can automate failed payment investigations. The goal is not just to resolve errors faster, but to prevent them from happening again.
As Trapani notes, this shift from reactive to predictive is where the real magic lies. What was once the exclusive domain of manual investigation is now supported by intelligent automation, freeing up human capital to focus on high-value decisionmaking. The result? Fewer delays, better audit trails, and happier customers.
This issue is filled with stories that echo and expand that theme of invisible intelligence.
In The Intelligent Bank, Jason Cao, CEO of Huawei’s Digital Finance BU, shares a bold vision for banking transformation. He outlines the “Four Zeros” strategy: Zero Downtime, Zero Touch, Zero Trust, and Zero Wait and he also introduces Huawei’s RAAS framework to support a new generation of resilient financial institutions. It’s a roadmap for the software-defined bank of the future, where uptime is constant, interactions are autonomous, and trust is embedded, not assumed.
Savas Manyasli, CTO at Finance Incorporated Limited, follows with a deeply thoughtful essay titled “Invisible by Design.”
Tawney Kruger, Editor, Financial IT
His premise is elegant: the best payment experiences are the ones we never notice. In an age of instant transactions across jurisdictions, currencies, and compliance frameworks, the challenge isn’t adding more features - it’s managing complexity behind the scenes. His vision of adaptive payment ecosystems is both philosophical and technical, with design simplicity as the highest goal.
Another urgent theme in this issue is identity and trust. Pedro Torres and Daniel Flowe sound the alarm on synthetic identity fraud, deepfakes, and the vulnerabilities of outdated KYC frameworks. Their recommendations include embracing decentralized identity (DID), integrating biometric authentication, and transitioning from static to dynamic identity validation. These reflect a growing consensus: our digital infrastructures must be as agile and intelligent as the threats they face.
Brady Harris, CEO of IXOPAY, shifts our attention from security to scalability in his sharp piece on payment orchestration. He makes a compelling case that for highgrowth businesses, sticking to one PSP (payment service provider) is no longer a viable strategy. Orchestration platforms not only optimize routing, reduce costs, and recover false declines - they futureproof your payment stack. In a landscape where decline rates exceed the cost of fraud, agility isn’t a nice-to-have; it’s a survival mechanism.
In the B2B sphere, Sarah-Jayne Martin of Quadient explores how AI and automation are finally bridging the gap between corporate and consumer payment
experiences. Her article highlights the importance of predictive analytics, intelligent reminders, and seamless approval flows in speeding up payment cycles and strengthening supplier relationships. With the rise of younger, tech-savvy financial decision-makers, this transformation is not just overdue - it’s expected.
Also featured: insights from Finastra’s Radha Suvarna on cloud and API-led modernization; a tour through embedded finance’s impact on credit unions by Union Credit’s Barry Kirby; and a spotlight on Huawei’s Dr. Peter Zhou, who discusses AIready storage as the foundation of financial data innovation.
All these voices reflect a common reality: financial technology is becoming less visible but more essential. It’s disappearing into the background, into APIs and automated flows, into decision engines and invisible fraud defenses. And in that invisibility lies power, not just technical but human. Power to deliver faster, fairer, and more intuitive experiences. Power to make finance work for people, not the other way around.
As I begin this chapter with Financial IT, I do so with gratitude, curiosity, and a commitment to clarity. I’m here to connect the dots between cutting-edge innovation and real-world value and to highlight the voices moving us toward a more intelligent, equitable financial future.
Let’s build it, one invisible transaction at a time.
Stay Curious,
Tawney Kruger Editor, Financial IT
Although Financial IT has made every effort to ensure the accuracy of this publication, neither it nor any contributor can accept any legal responsibility whatsoever for consequences that may arise from errors or omissions or any opinions or advice given. This publication is not a substitute for professional advice on a specific transaction.
No part of this publication may be reproduced, in whole or in part, without written permission from the publisher. Entire contents copyrighted. Financial IT is a Finnet Limited publication. ISSN 2050-9855
Yes, this is the future. This is where banks get supercharged. Quantum computing will enhance bank services by increasing speed and security. We are still a little time away from the Quantum impact. Traditional, banks have been slow to pick up the latest advancements. We always say that; “Banks want the latest and greatest, SECOND”. Banks are risk adverse and are not willing to take chances on tech they feel is unproven.
I am seeing a change in the way bank’s view progressive tech. There is an interest in
being in front of new technology. In the past, there was little interest until technology was established in the commercial world first. With Quantum, I do not know any banks that own a Quantum Computer. I have been told, by several banker buddies, that several banks are buying time on third party Quantum computers. This is an interesting development and approach for banks to do. First, the idea of using third-party anything is unusual. The approach of banks forming teams to work on unproven technology is revolutionary.
Why are these banks going after this differently than in the past. First the potential. Quantum computing promises dramatically increased speeds and the securest of security currently known. For banks this will speed up their networks of communication spanning Branches, ATMs, Payment Systems, Thirdparty providers, Governments, Corporations, Credit Bureaus, etc. This is an incredible amount of data and information that must move and move quickly. The thought is that Quantum Connections could provide speeds maybe four times or more. This benefit is clear. Security is a major benefit provided by Quantum. Today security is mainly cryptographically generated. This is what secures cryptocurrency. It is based on large complex math problems that have to be solved through computation. Current computing power is not able to solve these equations resulting in firm security. Quantum security is inherently different. Yes, it can process large equations faster due to the ability to run many calculations simultaneously. They can solve mathematical problems through Brute Force. Traditional computers can only do computations one after another in order. Sequential versus parallel is one of the prime differences between traditional computers and Quantum computers.
Quantum’s ability to solve multiple equations in parallel can be used to provide improved Cryptographic security. This is a difference maker, but where the real difference for security is layering with Lattice problems. Lattice problems don’t depend on factoring massive numbers as Cryptography does. Lattice-based security goes beyond factoring of prime numbers and hides a key in a high-dimensional lattice structure. The key hidden within the lattice is needed to be solved before the mathematical problem. Reveling of the key from such a lattice is considered likely impossible, even by futuristic Quantum machines.
Quantum computers will not replace your laptop. Yes, sorry but it is true. Quantum’s can’t help you stream movies faster, your games will not have better graphics, and your business apps will not cut hours off of your workday. No, Quantum will not improve your everyday tasks and in fact, would have the opposite effect. For now, traditional computers remain faster, cheaper and more reliable for everyday use and everyday tasks. This is especially true for financial transactions where exact and repeatable results are needed. What will happen is Quantum computers won’t
replace our computers but will complement them for specialized tasks.
The benefits are clear, speed and security, let’s take a look at the negatives. Instability, due to decoherence, of Quantum computers is an issue, likely solvable. Today they require extreme cooling or vacuum environments to insure stability. They require expensive infrastructure while your laptop is cheap, mass-produced, with optimized software and are widely available. Again, these are likely solvable issues.
Quantum issues that need more work are error correction, reliability and scalability. As these issues are continuously worked on, advancements must be made before Quantum computers are practical. Quantum computing is nascent technology. This potential can impact finance in fields like fraud detection, cybersecurity, supply chain logistics, risk analysis and financial modeling. Each of great importance to financial institutions. This explains why the early interest and beginning investment into Quantum computing.
Today, everything we hear about is artificial intelligence (AI). Everyone I talk to says Quantum and AI are two cutting-edge technologies that, when combined, will be amazing. When I hear someone say this, I think they don’t know enough about either. It is easy to sound good saying this, but is it meaningful?
What AI involves is training models to recognize patterns, make decisions, and to keep improving its responses. AI, based on Machine Learning techniques relies on large datasets and heavy computation. Quantum can enhance AI by speeding up and optimizing tasks. This is the potential to supercharge AI. Many claims about Quantum AI are exaggerated, which leads to misplaced investment when practical benefits are not realized. AI relies on large broad datasets while Quantum promises speedups of specific intensive linear operations. For the foreseeable future AI will remain on traditional computers where the AI algorithms are not able to benefit naturally from Quantum speedups.
The benefit of Quantum for the financial community will be a Hybrid Deployment. Traditional computers paired with Quantum computing. The majority of computing tasks will remain on traditional computers with specific types of functions doled out to Quantum computers.
The excitement of Quantum is real, only the future will tell us how real. I can only imagine the world where all of us are surfing on the Quantum Internet.
by Chris Principe, Publisher, Financial IT
EDITOR’S LETTER
Tawney Kruger, Editor,
John Trapani, Industry
Financial Services, Appian
PUBLISHER’S LETTER
Chris
Jason
Savas
FEATURED STORY
34 PRECISION VS. SCALE, OR WHY SMALL LANGUAGE MODELS ARE CRITICAL FOR HIGH-STAKES APPLICATIONS
John Byrne, Founder and CEO, Corlytics
INTERVIEW
Dr. Peter Zhou,
FEATURED STORY
Radha Suvarna,
26 SYNTHETIC ID FRAUD AND GENERATIVE AI: THE CASE FOR STRONGER IDENTITY FRAMEWORKS
Pedro Torres, CEO and Co-Founder, Youverse
28 REDEFINING THE FUTURE OF B2B PAYMENTS
Sarah-Jayne Martin, Director of Financial Automation, Quadient
30 BEYOND PASSWORDS: BUILDING CONTINUOUS TRUST IN THE AGE OF AI-DRIVEN FRAUD
Daniel Flowe, Head of Digital Identity, LSEG Risk Intelligence
32 HOW EMBEDDED FINANCE IS OUTPERFORMING PAID DIGITAL ADS IN CREDIT UNION GROWTH STRATEGIES
Barry Kirby, CRO and co-founder, Union Credit
36 BANKS ARE ALL IN ON REALTIME PAYMENTS – SO WHY ARE THEIR OWN FINANCIAL SYSTEMS STILL STUCK IN THE PAST?
Darren Heffernan, CEO, Trintech
37 THE ESSENTIAL ELEMENTS FOR TRUSTED TREASURY TRANSFORMATION
Morné Rossouw, Chief AI Officer, Kyriba
38 USING REAL-TIME DATA TO ENHANCE DECISION-MAKING
Mat Gapp, Head of Vision Next Product, Fiserv
42 CX IS THE NEW BATTLEGROUND FOR BANKS— AND INSTANT PAYMENTS ARE A FRONTLINE ADVANTAGE
Todd Clyde, CEO, Token.io
44 WHY PARTNERSHIPS AND PERSONALIZATION ARE THE FUTURE OF PAYMENTS
Simon Buchwaldt-Nissen, SVP, Head of Hospitality, Nexi Group
46 IT’S TIME TO TRANSFORM AFFORDABILITY ASSESSMENTS WITH AI
Zitah McMillan, CEO, Finexos
48 DEEP LEARNING, LEMS AND THE FUTURE OF QUANTITATIVE
Natahaniel Powell, CEO and founder, Deep MM
Brady
ARTIFICIAL INTELLIGENCE IN PAYMENT PROCESSING: EFFICIENT INVESTIGATIONS, HAPPIER CUSTOMERS
by John Trapani, Industry Lead, Financial Services, Appian
Artificial intelligence is one of the most impactful innovations the financial services industry has ever seen. From streamlining financial operations to enhancing customer experiences, artificial intelligence capabilities help financial sector organizations stay competitive in a marketplace that never stops shifting. The benefits of AI also extend to payment processes.
Why use artificial intelligence in payment processing?
Here’s a real-life example. A regulations technology organization focused on financial transaction screening wanted to help financial institutions more accurately
screen cross-border transactions for sanctions. This is a complex regulatory obligation and a costly barrier to moving to instant cross-border payments.
The organization found that 99.6% of the alerts generated on stopped payments were false positives. These false declines were frustrating their teams and negatively impacting customer satisfaction. They sought out technology to help support the investigation of matches and provide an outcome to sanction checks where autodisposition wasn’t possible.
They launched a new alert management tool with the help of an AI-powered automation platform that overlaid their existing financial systems to streamline workflows. This provided them with a fully
John Trapani is a Global Industry Leader at Appian Corporation, working with leaders in Financial Services to help drive growth and create opportunities for digital transformation using the Appian automation platform. In this role he works to ensure Apian delivers applications that help organizations grow, manage risk, and increase the efficiency and effectiveness of their business processes.
integrated, configurable user experience for managing and working with alerts across multiple teams and levels of investigation through to a standardized outcome.
Their new alert tool enhanced the user experience and minimized steps to improve investigation productivity. It also helped them store and retrieve evidence while complying with agreed standards to eliminate having to rework investigations and improve audit trails. As an added benefit, the platform gave the organization the flexibility to more easily implement new regulatory requirements, ensuring they were prepared for the future.
Moving beyond the limitations of traditional systems
Traditionally, conducting payment investigations, resolving failed transactions, and answering common customer queries have been resourceintensive processes. This is particularly true when organizations rely on outdated legacy systems, which struggle to adapt to evolving business requirements.
Using AI in the payment process eases the burden on payment providers by letting advanced data analysis, predictive analytics, and automation tools take over some of the work.
AI addresses common pain points in the payments industry, minimizes human intervention, and automates routine tasks. AI algorithms can reduce false positives and accurately identify cases that need further investigation. And capabilities like agentic AI support customer assistance, answering customer queries and providing financial advice based on market data. All of this leads to a better customer experience with faster processing times and less payment friction.
Here are four ways AI is changing the payment industry.
1. Faster payment investigations
Payment investigations are a timeconsuming and error-prone process. When there are payment delays, inefficient payment routing, or payment failures, banks and payment service
providers must sift through huge amounts of transaction data to set things right. The manual tasks involved lead to long resolution times, high costs, and erosion of customer relationships.
With AI, you can automate the investigation lifecycle to reduce payment delays. AI-powered payment solutions can analyze huge amounts of payment data in seconds, identify anomalies in transaction data, and guide operations teams to avoid potential risks. Machine learning models, for instance, can use predictive analysis to remediate failed transactions and offer recommendations to prevent future errors. These benefits not only expedite investigations but also significantly reduce operational costs.
2. Better customer experience
A satisfying payment experience is critical for customer loyalty. AI-driven solutions help organizations in the financial services industry deliver faster, more accurate responses to customer assistance queries. Chatbots and virtual assistance use natural language processing (NLP) to handle routine inquiries 24/7, taking on basic tasks like checking transaction history or providing details about cross-border payment fees.
For more complex cases, AI-driven tools augment human agents by giving them real-time analytics with suggested resolutions. This hybrid, mixedautonomy approach that combines AI with human oversight means customers receive efficient service yet can still get personalized support when a human brain is needed. Plus, customer service agents get to spend more time on higher value work.
3. Improved fraud detection
Payment providers and systems are frequent targets for fraudulent activity. Robust cyber fraud detection and prevention mechanisms are needed to safeguard customer funds. AI-powered fraud detection can monitor transactions in real time and use predictive analytics to identify suspicious activities and anomalies in transaction data with
greater accuracy than traditional rulebased methods.
By flagging potentially fraudulent transactions as they occur, financial institutions can protect their customers and minimize risks. Because of AI's deep learning capabilities, financial institutions can continuously adapt to new fraud techniques. This ensures fraud processes remain effective as potential fraud prevention patterns evolve.
4.
Savings on operational expenses
AI-driven automation in payment processing offers substantial cost savings. Companies that leverage AI and automated workflows reduce the need for human intervention, allowing financial institutions to allocate resources toward other strategic activities. AI-driven payment solutions can also improve regulatory compliance, analyzing transaction data so banks can avoid penalties and financial losses.
In addition, faster and more accurate payment investigations mean less revenue leakage. These efficiencies translate to better service delivery, creating a win-win situation for organizations in the financial industry and their customers.
Embracing the future of the payments industry
Using artificial intelligence in payment processing is no longer optional. It is a strategic imperative for financial institutions to remain competitive and improve long-term customer trust. By leveraging artificial intelligence capabilities for payment investigations, customer engagement, fraud processes, and operational efficiency, banks and payment service providers create a more resilient and customer-centric payments system.
As the financial industry continues to shift toward a digital-first future, those that embrace AI-driven payment solutions will win out. Efficient digital payment systems lead to happier customers, and in the fast-paced world of finance, customer satisfaction is the ultimate currency.
Jason Cao,
CEO of Digital Finance BU, Huawei
Jason Cao is the CEO of the Huawei Digital Finance. He manages Huawei's business in the global financial industry. The Digital Finance Team strives to meet the needs of financial industry customers through ongoing innovation, works with partners to develop leading solutions, and aims to shape smarter and greener finance through full connectivity and intelligence.
THE INTELLIGENT BANK: A BLUEPRINT FOR RESILIENT, AI-DRIVEN FINANCE
In an exclusive interview with Financial IT, Huawei’s Jason Cao outlines how AI, autonomy, and the 4-Zero strategy are driving the transformation toward intelligent, resilient banking.
The Mobile World Congress is one of the largest events of the year. In 2025, MWC hosted around 110,000 participants in Barcelona, Spain. The MWC is one of the key technology events each year. Huawei, one of the world’s largest multinational technology companies, showcased an impressive display of tech expertise. Huawei, from their key position in Hall 1 of the 8 event halls, put on the greatest display of innovation for a range of industries. Huawei’s financial solutions showed impressive leadership in their breadth and depth of applications.
During MWC 2025, Jason Cao, CEO of Digital Finance BU, shared his vision about the Intelligent Bank, which Huawei sees as the future. As the leader of Huawei’s financial solutions, Jason Cao has driven a broad approach focused on banks and financial institutions and believes that a major key for the future of banking is resilience.
Cao spoke about the evolution of the bank. Specifically, he cited that the transactional bank became the digital bank, which will now evolve into the intelligent bank. Huawei aims to drive this evolution by focusing on resilience. Cao believes that by improving resiliency, they can enable non-stop banking. This will be evaluated through an approach Cao terms "Bank 4 Zeros."
Cao breaks down Bank 4 Zeros in the following framework:
1. Zero Downtime: Real-time payments have dramatically increased and are difficult for traditional platforms to support. Eliminating downtime is imperative for the Intelligent Bank to meet its clients' requirements.
2. Zero Touch: A vision of providing clients with greater autonomy to reduce or even eliminate human intervention. Using AI models for autonomous driving as inspiration, it is possible to visualize changes to networks, traffic, and applications in real time without needing the human touch.
3. Zero Trust: Today, banks face increased vulnerabilities at every connection and each data point. Threats and attacks come from various sources. Huawei provides security solutions that provide the trust, protection, and security that banks of today demand and require.
4. Zero Wait: The customer experience is paramount to the Intelligent Bank. AI can be utilized to make real-time recommendations for clients, the bank, and its personnel. The strength of AI is in analyzing data to provide immediate results.
In addition to the four zeros, Huawei introduced its AI-powered R-A-A-S (Reliability, Availability, Autonomy, and Security) framework at MWC. This framework is designed to help banks have extremely resilient infrastructure. Huawei believes that this end-to-end framework of RAAS and the 4 Zeros will enable a new generation of Intelligent Banking.
Cao also took the time to address some of the challenges that exist in implementing these frameworks and enabling the rise of the Intelligent Bank.
One of the main issues Cao cites is that legacy systems make it difficult to evolve. "We are speaking to many banks and all the
different departments inside banks. The biggest inhibitor for banks is their legacy systems. This makes a bank's ability to transform very difficult."
Despite this challenge, Cao emphasized that Huawei has a "step-by-step" approach that it has experience implementing with several big banks that have successfully transformed using Huawei's frameworks and expertise. Cao noted that "the future is for the software-defined bank to serve the digital human," which is why this process is beneficial despite challenges.
Finally, Cao touched on how these efforts are enabled by a top-down vision throughout Huawei that provides a vision and the resources to create The Intelligent Bank. Cao, a firm believer in the future, emphasized that the "important part is the quality of our people." He cited the "experience, ingenuity, and imagination" that the team at Huawei possess as they seek to push the boundaries of solutions for banks and financial institutions.
INVISIBLE BY DESIGN: THE DISAPPEARING ACT OF PAYMENTS
The invisible complexity of simplicity
Mark Weiser, former CTO of Xerox PARC and a visionary in our field, once said: "The most profound technologies are those that disappear. They weave themselves into the fabric of everyday life until they are indistinguishable from it." As a fellow CTO, that philosophy resonates deeply with me. At FIL, we aim to create technologies that feel intuitive, unobtrusive, and ultimately essential – not because they shout for attention, but because they simply work.
The future of payments lies not in their prominence, but in their disappearance from conscious thought. The best payment experience is one that users barely notice. They are seamless, instant, and require minimal effort. Yet achieving this apparent simplicity requires navigating tremendous complexity behind the scenes.
The tale of two seconds
Consider a seemingly straightforward scenario: Jane lives in London, earns her income and pays tax there. While visiting a market in Shenzhen, China, she spots a high-tech camera she's always wanted at an attractive price. On the shop door, she sees a dozen contemporary payment methods and a sign reading "Duty Free Tax Exempt/Return." She pays using the shopkeeper's handheld device, enters her UK VAT number, and the transaction is complete in seconds.
A simple transaction, right? Just currency conversion, a few digital debits and credits, and we're done.
But let's peek behind the curtain at what actually transpired in those two seconds:
• A payment gateway certified by multiple agents and operating under strict PCI-
DSS standards processed the transaction
• An FX conversion occurred between financial institutions, often mediated by regulatory authorities
• Fraud and identity checks verified that Jane was spending her own money and cleared her from an anti-money laundering perspective
• GDPR compliance ensured that Jane's bank shared only the necessary personal information
• Jane's bank settled and reconciled with their Chinese counterpart
• The system calculated VAT using Jane's tax number and communicated with UK tax authorities via their API service
This simple transaction triggered at least a hundred database operations and a dozen network messages across multiple jurisdictions and regulatory frameworks. The technological infrastructure supporting this moment represents years of experience, thousands of days of development, and countless hours of problem-solving.
The Cartesian complexity of cross-border finance
The regulatory landscape governing trade, taxation, and money transfers contains multitudes of laws and rules, some with roots dating back to Roman times. Even in our simple cross-border scenario, the Cartesian combination of these variables creates a complexity so overwhelming that it could keep the financial technology industry busy for decades.
Adding to this challenge, these rules can change dramatically due to geopolitical shifts and economic fluctuations. What works today might require significant reconfiguration tomorrow.
Despite these challenges, technology has made remarkable progress. The financial technology sector has an inherent drive toward standardization. Enterprises, governments, and supranational organizations such as the EU, ISO, SWIFT, EPC, and NACHA in the United States have established and improved standards, created frameworks for communications protocols and messaging, and initiated paradigm shifts like Open Banking. Many fundamental issues have been addressed and resolved. Financial technology is often ahead of the curve in communication and messaging protocols, thanks to the unifying efforts of various stakeholders. Cross-border regulations may sometimes feel like hurdles, but they also push the ecosystem forward, ensuring fairness, security, and transparency.
Simplicity as the ultimate sophistication
At FIL, we've developed a straightforward test for evaluating our innovations: "Can you explain it to your grandfather in one sentence?" This principle guides our approach to simplifying the complex world of financial technology.
We aim to create payment experiences that don't burden users with the intricacies of regulations, protocols, and security measures happening behind the scenes. Our focus on simplicity shapes our vision in several ways:
Designing for human interaction
We aspire to develop payment systems by starting with the human experience rather than technical capabilities. Our goal is to create interfaces that feel natural and intuitive regardless of a user's technical literacy or cultural background.
Savas Manyasli, Chief Technology Officer, Finance Incorporated Limited (FIL)
Savas holds a B.Sc. in Applied Mathematics and brings over 20 years of software development experience across sectors including manufacturing, telecoms, and financial services. Since 2001, he has held senior roles in enterprise computing, software, banking IT, leading core system design, infrastructure, vendor management, and IT governance. As Chief Technology Officer at FIL, Savas oversees all technology operations and development. He plays a key role in shaping and executing the company’s IT strategy, ensuring systems are built from the ground up to support FIL’s product and service delivery.
Building invisible complexity
We aim to keep the most sophisticated aspects of our systems hidden from view. By thoughtfully managing complexity behind the scenes, we intend to present users with clean, straightforward interfaces while continuously enhancing the underlying technology.
Creating adaptive ecosystems
Rather than requiring users to navigate different systems for different payment scenarios, we're working toward building frameworks that adjust seamlessly to context. Our vision is for the underlying technology to reconfigure itself based on the situation while maintaining consistent interaction patterns.
We believe the future of payment simplicity isn't about reducing functionality. It's about thoughtfully managing complexity so users can accomplish their goals with minimal effort.
The tech horizon
As we look to the future, several emerging technologies may help further simplify global payments:
Unified authentication approaches
The proliferation of authentication methods can create complexity for users. We're exploring solutions that will provide strong security while reducing the cognitive load associated with authentication. Biometric verification, behavioural analysis, and contextual authentication could all be part of this evolution.
Compliance-aware processing
We're interested in approaches that could integrate regulatory compliance more seamlessly into payment flows. Such approaches should reduce friction
while maintaining or enhancing security by enabling appropriate interventions when necessary.
Cross-border consistency
While complete global standardization of financial regulations seems unlikely, we hope technology can help bridge some gaps. We aim to design systems that can abstract certain complexities of cross-border transactions, working toward more consistent experiences regardless of location.
The human element
Despite all the technological sophistication, we never lose sight of the human element. The best payment systems are those that adapt to human behaviour rather than forcing humans to adapt to technology.
We know that our technologies, and the technologies we interface with, are inexorably moving towards being virtually instant in terms of time. But this is not our primary focus. We believe that the future of payments is about making financial interactions more natural, more secure, and more accessible to everyone. It's about creating technologies that disappear into the background, enabling people to focus on what truly matters: living their lives, running their businesses, and connecting with others.
As we continue to navigate the complex intersection of technology, finance, and human behaviour, our north star remains clear: to inspire the financial services industry to better serve humanity. And in the realm of payments, that means creating experiences that are truly instant, frictionless, and global – not because they're technologically impressive, but because they simply work.
That's the future we're building at FIL, one invisible transaction at a time.
HOW TO KNOW IF YOU’RE READY FOR PAYMENT ORCHESTRATION
When I talk to merchants around the world, I often hear this question: “How do we know if we’re ready for payment orchestration?”
It’s the right question to be asking.
Today’s payments environment is more dynamic – and more complex – than ever before. Businesses are operating across multiple regions, channels, and customer types. And while that kind of growth creates opportunity, it also creates friction in your payment stack. Operational inefficiencies, rising fees, high decline rates, and fragmented integrations aren’t just a nuisance anymore – they’re revenue blockers.
This is where payment orchestration enters the picture.
At its core, payment orchestration is about managing and optimizing multiple payment providers, gateways, and methods through a single platform. But it’s more
than just infrastructure – it’s a strategy that gives you control over how payments move through your business.
So, how do you know when you’ve reached the tipping point? From what I’ve seen, a few common patterns suggest it’s time to take the leap.
Cost Pressures Are Mounting
If you’re locked into one payment processor, you’re likely paying more than necessary. Every provider has a different fee structure, and when you’re limited to a single option, you lose pricing leverage and flexibility.
Orchestration allows you to route transactions based on cost, geography, or performance in real-time based on rules you set in place. That level of control can result in significant savings – especially as your transaction volume grows. It also
Brady Harris, Chief Executive Officer, IXOPAY
Brady Harris, a visionary FinTech Executive, has over two decades of experience leading high-growth financial technology and SaaS companies. As the CEO appointed following the merger of IXOPAY with TokenEx, he is dedicated to advancing global payment solutions. Harris employs a hands-on leadership style, focusing on operational excellence and strategic growth initiatives. During his tenure at Dwolla and Payscape, he oversaw significant increases in payment volumes and user engagement and fostered cultures that prioritize agility, high performance, and alignment with core values. Additionally, his expertise in mergers, acquisitions, and scaling companies to successful exits is key as IXOPAY positions itself to become the one-stop payment industry solution. Harris's commitment is not only to enhance enterprise value but also to empower his teams and clients in the evolving digital payments landscape.
gives you the agility to switch providers without long lead times or added complexity.
Decline Rates Are Cutting Into Revenue
False declines are more than a customer experience issue – they’re a massive revenue leak. According to Riskified , false declines cost businesses $430 billion annually, exceeding the entire decade-long projected global loss to card fraud, which McKinsey estimates at $400 billion. In other words, the cost of saying "no" to good customers has quietly surpassed the cost of actual fraud.
One of the most impactful benefits of payment orchestration is the ability to intelligently route or retry soft declines across multiple providers. You can recover revenue you would have otherwise lost, and your customers get a smoother experience. Everybody wins.
You’re Expanding into New Markets
Global expansion is an exciting step, but it brings its own challenges: multiple currencies, regional regulations, and local payment preferences. Trying to handle all of that manually – especially with a legacy PSP integration – can slow your time-tomarket and inflate your overhead.
Orchestration platforms connect you to a wide network of regional and global payment providers, allowing you to turn on new capabilities or methods with minimal effort. Whether you’re entering Brazil and need PIX or supporting SEPA in Europe, the right platform will have the integrations ready to go.
Payment Flexibility Is a MustHave
Consumer expectations have evolved. They don’t just want options – they expect them. Credit cards, wallets, bank transfers, BNPL… if you don’t offer the right mix, you risk losing the sale.
Orchestration platforms make it easy to add and manage multiple payment methods from one interface. That means you can tailor your checkout to customer preferences without building and maintaining a dozen separate integrations. It’s good for conversions, and great for your development team.
Fraud and Chargebacks Are Growing Concerns
Fraud is always evolving. And when you’re relying on a single fraud engine, you’re often forced to choose between being overly cautious or overly permissive.
With payment orchestration, you don’t have to choose. You can integrate multiple best-in-class fraud providers for specific use cases – whether it’s behavioral profiling, policy abuse detection, or chargeback guarantees – and route transactions to the provider best suited to assess the risk.
This layered, flexible approach means stronger protection and fewer false positives, which translates to less fraud and higher approval rates.
Integration Overhead Is Draining Resources
If your development team is spending more time maintaining payment integrations than innovating, that’s a red flag. Multiple dashboards, mismatched APIs, manual reconciliation – it all adds up.
Orchestration simplifies this by consolidating all your providers into a single API and dashboard. You gain visibility, streamline workflows, and significantly reduce integration maintenance. That’s not just more efficient – it’s more scalable.
You’ve Hit a Ceiling
Growth shouldn’t be held back by your payment infrastructure. But I’ve seen too many companies stuck with brittle, outdated setups that can’t adapt. Every time they try to scale, they run into outages, contractual limitations, or other bottlenecks.
Payment orchestration breaks those constraints. It gives you the agility to adapt your payment stack in real-time, onboard new providers faster, and ensure your payment operations scale with your business – not against it.
The Bottom Line
If any of this sounds familiar, you’re not alone. Most businesses outgrow their payment setups faster than they expect. And while orchestration may sound like a big shift, the truth is that it simplifies things – not just for your tech stack, but for your entire organization.
At IXOPAY, we’ve helped hundreds of growing merchants modernize their approach to payments with orchestration. The results speak for themselves: lower costs, fewer declines, faster market launches, and better customer experiences.
The question isn’t whether you’re ready. The question is whether your current payment stack is ready for where your business is going next.
In an exclusive interview with Financial IT, Huawei’s Dr. Peter Zhou discusses how AI-ready storage solutions are reshaping the financial industry’s data landscape.
One of the largest and most important events every year is the Mobile World Conference held in Barcelona, Spain. Huawei Technologies had the largest presence of any company at the MWC. Huawei is one of the largest multinational technology companies in the world. Financial IT was represented by Publisher Chris Principe. Financial IT had the opportunity to sit down with Huawei's President of Data Storage, Dr. Peter Zhou.
Chris Principe, Financial IT: Dr. Zhou, thank you for spending time with Financial IT at MWC. You have been a leader in digital transformation. What do you see as some of the keys in the storage field?
Dr. Peter Zhou: Thank you, Chris, Huawei, as well as myself are pleased to speak with Financial IT. To be AI-ready, banks need to engage in the process of transforming data into information and knowledge. Banks must integrate data storage, data management, and resource management with AI to speed up modelling and to empower their enterprises.
Financial IT: Tell me more about a bank's storage needs as AI usage increases.
Dr. Peter Zhou: The reality of AI innovation will enable banks to enhance their operations, processes and client engagement in ways we have not seen before. Storage is a key element in making this transformation through AI a reality. Banks need to understand what storage options meet their internal needs, client demands, and security requirements. This means that banks must look at the different options, such as hard disk drives (HDDs), solid state drives, and cloud storage or likely a combination of the three.
Chris Principe: That is quite interesting, and I am fascinated that you consider hard disk drives as part of the solution for banks. I thought that hard disk drives were old technology and outdated compared to solid state drives and cloud storage.
Dr. Peter Zhou: Yes, Chris, I think many people think that HDD are a technology of the past. Today, there is still an important role for HDD technology, especially for banks. It is true that HDD are slower, have
higher consumption and are at risk of being damaged. In a bank setting, these factors are not considered risks or drawbacks. The benefits of HDD over SDD in a bank environment are the higher capacities that are available and the lower cost per gigabyte (GB). HDDs outperform SSDs for cold storage, mass storage and heavy workloads.
Financial IT: These are important differences, and I can see the value in each of the three types of storage: HDD, SDD, and Cloud. How do you see Huawei storage meeting the strategies that banks have today?
Dr. Peter Zhou: Good question, and let me first point out that AI is impacting storage through capacity and resource management. Digitalization facilitated through AI enhances the process of transforming data into information and knowledge. Data storage needs AI innovation to refine performance, capacity, and resilience. AI technologies can significantly boost cluster efficiency, accelerate performance, reduce latency, and elevate the user experience. Banks require the deployment of large-model storage applications. At Huawei, we have vast experience in providing products and software as an integrated solution for the needs of banks and financial institutions.
Financial IT: In the financial community there are a variety of needs in the use of storage, how do you look to meet the challenges that exist?
Dr. Peter Zhou: Within Huawei, we see that a combination of storage products is required to meet the usage needs that banks have. First, let’s look at the types of data requirements that banks have. I will put it in three categories:
• Cold Data – Archival – Hard Disk Drive
• Warm Data – Transactional – Solid State Drive
• Hot Data – User Actions – Cloud Storage
These are useful generalizations that broadly cover the data usage of banks today. The game changer for data storage is the deployment of AI to optimize the data
storage methods and the storage processes. At Huawei, we have successfully deployed our products and solutions with several larger banks. Our bank clients have shown quantifiable benefits that optimize how storage is utilized. Huawei Data Storage provides diverse data storage services, empowering banks to turn their data into assets to reveal the value in their data. For banks, it is not about gold in their vaults, it's about the gold in their data storage platforms.
Financial IT: Peter, one of the things that I find interesting is Huawei’s ability to be in front of the industry with technology that leads the industry.
Dr. Peter Zhou: Yes, Chris, for Huawei, it is a consistent effort that is focused on advancement, innovation and investment. Let me say that Huawei’s commitment to building AI-ready data storage is combined with a vision of future-proof storage based on AI for the financial services industry. AI makes data more important, and when data is ready, it becomes an asset. Looking ahead, we will keep promoting a strategy of using various data storage approaches combined with AI to strengthen our collaboration with banks and financial institutions.
Financial IT: Clearly, the future is full of excitement, and Financial IT will be interested to watch as Huawei brings us a better future through technology. Thank you, Dr. Zhou.
A Product and Business leader with over 27 years of diversified experience across Payments, Cards, and Consumer Lending for Consumer and Commercial client segments. Built and transformed various functions and businesses across multiple geographies including U.S. and Asia.
Currently the Chief Product Officer for Payments at Finastra, responsible for the overall strategy, execution and build-out of payment processing engines and Financial messaging solutions.
Prior to this, Radha was the EVP/Head of Enterprise Payments for Citizens bank, with responsibility for Enterprise Payment platforms, Consumer payment products, Treasury payments strategy, and Payments innovation. During this time, Radha was the Chair of Secure Token Exchange board at The Clearing House (TCH). Before joining Citizens, Radha spent over 20 years at Citi. He was the head of Digital Payments & Lending for US Consumer bank at Citigroup, with responsibilities for end-to-end mobile/ web products for money movement, digital commerce, unsecured loans, and associated industry partnerships with big-tech and fintechs.
Radha held number of diversified roles prior to that, including P&L ownership for Credit Cards/Loans/Merchant Acquiring portfolio for Citibank Singapore, Citi Asia Network relationship lead with Visa/ MC, and Credit Risk Manager at Citibank US.
ACCELERATING PAYMENTS MODERNIZATION
Financial institutions are showing no sign of slowing down when it comes to deploying new technologies and improving their capabilities. That’s one of the key findings of Finastra’s Financial Services State of the Nation Survey Some 98% of financial institutions surveyed globally said they are actively taking steps toward change, recognizing the urgency of preparing for technological disruption.
Modernizing payments
A key area where financial institutions know they need to invest is in payment modernization. The ever-increasing volume of payments that banks are required to process, and the adoption of instant payments, require payment systems to be able to scale rapidly and be highly available.
A 2024 report from Omdia on ‘Embracing the revenue-generating opportunities of a modern payment hub’ found that nearly two-thirds of banks’ IT budgets is spent on maintaining existing legacy technology, and just 36% is allocated to either growing or transforming their technology. But the regulatory drive behind real-time payments, the implementation of the ISO 20022 standard, and open banking maturity mean that banks are increasingly being pushed to upgrade payment systems to fit with new market infrastructure.
Financial institutions need a strong technological foundation to ensure they are ready for change, and able to adapt beyond legacy systems. This foundation is built on two critical components: cloud solutions and API-led technology. These technologies are not just facilitators; they are the backbone of a secure, resilient, and adaptable financial future. According to our research, close to half of decision makers (47%) are modernizing all aspects of their institutions’ operations.
By moving to the cloud, financial institutions can store and process vast amounts of data quickly and securely, benefitting from greater scalability, flexibility, and cost-efficiency. Unsurprisingly in this context, a quarter (27%) of financial
institutions have improved or deployed cloud solutions in the last 12 months.
APIs allow for the seamless integration of financial services, creating a more interconnected and efficient financial ecosystem. This interoperability is key to fostering innovation, as it allows financial institutions and third-party providers to collaborate and build new, customer-centric services. Similar to cloud adoption, a quarter of institutions (27%) have deployed or improved open APIs in the past 12 months, further highlighting the effort to modernize operations in an evolving sector.
Solutions built with an open API mindset enable banks and other institutions to simplify their integration with other systems, adapt to emerging payment trends and to scale in line with the demand for the new business functionality being developed at a rapid rate. Containerized, microservices-based architectures are helping banks respond faster, personalize and differentiate their financial services, enhance customer experiences and unlock revenue.
Democratizing access
Many financial services are not accessible to underserved communities or small businesses which hinders financial inclusion. Solutions underpinned by the cloud and open APIs allow for easy integration and collaboration between fintechs and financial institutions, making it simpler to offer tailored financial services to broader market segments. This includes more accessible, transparent and cost-effective cross-border payments solutions available through an open ecosystem of specialist fintechs and cross-border payment partners. Similarly, cloud-based solutions are also helping to level the playing field for institutions of different sizes by empowering lower-tier banks to make finance more accessible to their customers. Collectively these developments are helping more individuals and businesses to access the modern payments infrastructure they need to thrive, fostering a more inclusive financial ecosystem.
Enabling technologies
When asked about the uptake of enabling technologies, financial institutions are prioritizing investment in technologies such as Generative AI (Gen AI), embedded finance, and Banking as a Service.
Uptake in AI is accelerating rapidly. Around two in three financial institutions (61%) have deployed or improved their capabilities in AI in the last 12 months, up from 37% in 2023.
The proportion of financial institutions that have adopted or improved their Gen AI capabilities increased more than for any other technology over the year, with 35% of financial institutions across markets saying they have done so (compared to 25% in 2023). It’s evident that financial institutions understand that the sector is now one defined by rapid technological transformation, and that slowing down could mean falling behind.
Last year’s research highlighted the perceived significance of Gen AI in personalizing customer offerings. Almost all (97%) of financial institutions are now offering personalized services, recognizing that they are an expectation, not just a differentiator.
Real-time payments are the most commonly offered personalized service across the majority of countries surveyed. This differed in the US market where customizable digital wallets that allow for personalized payment options and transaction categories was cited as the most widely offered service. The use of chatbots to accurately answer customer questions and provide 24x7 customer support also featured highly.
It’s clear that novel payment practices, coupled by technological advances are transforming how services are delivered, making them more accessible, personalized, and efficient. These advancements are better enabled through a secure, interoperable framework for data sharing and integration and in enabling financial institutions to build an innovative environment to keep pace with change.
Pedro Torres, CEO and Co-Founder, Youverse
With many years of experience in technology, innovation and product design, from large telco companies to medium-size companies and start-ups, and a strong research background, Pedro Torres has worked for the last many years in global executive positions in customer experiences based on biometrics to enable immersive and seamless journeys. A strong believer in decentralized approaches to privacy, Pedro has been leading efforts to provide the necessary protection and control to users as they authenticate for convenience with their face in multiple services such as proving identity to create a bank account, automatically checking-in to hotels, picking up car rentals or pay for goods and use loyalty in a fully contactless on-the-move fashion.
Deepfake fraud surged dramatically, with incidents skyrocketing in the last few years. This alarming rise underscores a significant shift in the landscape of identity fraud, propelled by advancements in generative AI. What once required social engineering and stolen documents can now be achieved with a few lines of code and a generative AI model.
Unlike traditional identity theft, where criminals hijack an existing identity, synthetic identity fraud involves the fabrication of entirely new personas. These identities are built by blending fragments of real data, like a stolen national ID number, with invented information, creating digital “ghosts” that can be used to open accounts, build credit histories, and conduct fraudulent transactions over extended periods without raising immediate suspicion.
Generative AI has become a critical enabler for fraudsters, offering new ways to forge credibility in digital spaces, generating fake documents that closely resemble legitimate IDs or passports; creating deepfake videos that impersonate real users; and producing fabricated biometric data capable of bypassing weak liveness detection systems.
Synthetic identities are especially appealing to cybercriminals because they are longlasting, easy to replicate at scale, and often fly under the radar of traditional fraud detection tools. Once established, a synthetic identity can operate like a real person, opening bank accounts, applying for loans, even passing KYC checks.
According to the U.S. Government Accountability Office, synthetic identity fraud now represents over 80% of all new account fraud, a staggering figure that underscores how ill-equipped existing systems are to counter these attacks.
The cracks in current identity verification systems
Despite the growing sophistication of identity fraud, most verification systems remain stuck in the past, relying on static data points and outdated assumptions about trust. At the core of many onboarding and KYC processes are elements like government-issued IDs, credit history, and knowledge-based authentication. But these were designed for a time before AI could fabricate IDs and simulate biometric data. As a result, the tools meant to keep fraudsters out are increasingly easy to manipulate.
Even biometric systems, once considered foolproof, are not immune. Many rely solely on
facial authentication without robust liveness detection or multi-modal verification.
Speed is often prioritised over security. In a bid to reduce friction and win users quickly, financial institutions have been streamlining identity checks, accidentally opening the door to synthetic profiles that slip through with minimal resistance. The result? A growing gap between the threats banks face and the defences they rely on. And as fraudsters innovate with generative tools, every static identity check becomes a potential liability.
Why we need stronger identity frameworks
Identity is no longer just a checkbox at onboarding, it’s the backbone of digital trust. And right now, that trust is eroding.
Legacy identity frameworks depend on centralized databases, static credentials, and outdated verification logic. These systems assume that what was once true remains true indefinitely. But synthetic identities break that assumption by introducing users who are entirely manufactured yet carry credentials that seem authentic.
The problem isn’t just technological, it’s structural. Centralized identity systems create single points of failure: if a database is compromised, thousands or millions of identities are at risk. We need frameworks that can adapt to the fluid, high-risk digital environments we operate in today. Systems that don’t just verify an identity once and move on, but validate it continuously, contextually, and cryptographically.
Stronger identity frameworks should be:
• Decentralized – shifting control from central authorities to the user, reducing attack surfaces and preventing large-scale breaches.
• Verifiable by design – so every credential carries proof of issuance and can be independently confirmed.
• Privacy-preserving – ensuring users can prove who they are without oversharing sensitive information.
Decentralized identity (DID) frameworks allow individuals to store and manage their own identity credentials securely while giving organisations cryptographic proof that those credentials are valid. Instead of trusting the document, you trust the issuer and the verifiability of the claim.This approach reduces reliance on honeypot databases and gives users more control over their data. If one credential is compromised, the rest remain secure.
Biometric authentication, when implemented responsibly, offers a level of security that’s hard to fake. Facial authentication can serve as an high-assurance anchor to real, living individuals. But the key is multi-modal, AI-aware biometrics, with a layered approach that includes liveness detection, movement analysis, and sometimes even passive behavioural signals.
When combined, DID + biometrics create a powerful defence mechanism:
• DID provides the verifiable credential history.
• Biometrics ensure the person presenting those credentials is real and present.
Together, they make it vastly more difficult for synthetic identities to enter or persist in a system unnoticed.
Every day that organizations delay updating their identity frameworks, the cost of inaction grows. The financial damage is measurable: losses from synthetic identity fraud are projected to run into billions annually. But the hidden costs are often worse: damaged reputations, broken user trust, and regulatory fallout.
Moreover, regulators are catching up. Financial institutions will increasingly be expected to demonstrate not just compliance, but proactivity: proof that they’re investing in the right technologies to detect and prevent identity manipulation. Inaction now could mean fines, sanctions, and becoming a case study in what went wrong.
While institutions wait for clearer rules or safer ground, threat actors are already steps ahead, leveraging technology in ways that legacy systems were never designed to withstand. Generative AI isn’t the villain, it’s a tool. But like all tools, it reflects the intentions of those who wield it. And right now, fraudsters are wielding it faster and more effectively than many institutions can respond.
To fight synthetic identity fraud, we need identity frameworks that are just as intelligent, adaptive, and scalable as the threats they face. That means embracing decentralized identity, investing in AIresistant biometric authentication, and abandoning legacy systems that treat identity as a static checkbox. The choice is clear: evolve, or be outsmarted.
Sarah-Jayne Martin, Director of Financial Automation, Quadient
Sarah-Jayne Martin holds over 25-years of experience in finance, reviewing current processes, identifying inefficiencies, and executing solutions. Having managed global teams and held stakeholder and project manager positions for various payment related global initiatives, she holds a wealth of knowledge of the payments industry. Sarah-Jayne frequently takes part in thoughtleadership event panels, webinars and conferences, continuing her commitment to continuously expand and sharpen her skillset.
REDEFINING THE FUTURE OF B2B PAYMENTS
The business-to-business (B2B) payment space is in the middle of a digital revolution. Organisations are reimagining how they move money, through innovations such as AI, mobile wallets and embedded finance. As expectations rise, there is increasing pressure for B2B transactions to match the ease and speed of business-to-consumer (B2C) payments. For example, instant contactless payments that many consumers are now accustomed to.
A key driver behind this shift is the influx of younger, tech-savvy decision-makers in the B2B world. This generation seek greater agility and efficiency to match their dayto-day life when it comes to fast payments. Alongside these developments, the Council of the European Union announced the adoption of new regulation on instant payments last year. This regulation mandates that instant Euro payments be fully accessible to both consumers and businesses across the EU and EEA countries. With such a strong move towards real-time financial infrastructure, it’s only a matter of time before neighbouring countries like the UK align.
Friction in traditional B2B payment systems
Despite rapid advances in consumer payment tech, B2B transactions have lagged behind. Many businesses still rely on outdated methods like physical cheques and delayed bank transfers. These dated practices not only damage cash flow but also strain supplier relationships and slow down service delivery. Yet, research shows that 70% of businesses report having no plans to discontinue cheque usage in the next two years.
Even though digital options such as direct bank transfers are widely available, many companies have yet to fully embrace them. One reason for the slow adoption is limited use of AI and intelligent automation – technology that combines AI with automation to enable systems to make decisions and execute tasks without human input.
The lagging implementation of these innovative tools creates several challenges. Without automation, payment approvals and settlements can take days, causing bottlenecks and missed opportunities. At the same time, AI-powered tools that can provide predictive insights and realtime anomaly detection are often absent, leaving organisations reactive rather than proactive. The lack of transparency across fragmented systems further complicates tracking payments, increasing confusion and the likelihood of financial reporting errors. Moreover, without AI-driven reminders or follow-ups, organisations risk missing due dates, increasing the potential for late or defaulted payments.
Harnessing AI and automation for real-time B2B payments
To address these issues and meet rising expectations, businesses must modernise their finance departments. However, simply shifting from paper-based processes to online portals isn’t enough. Instead, embracing AI-powered tools and intelligent automation is crucial to enable real-time payments in the B2B space.
AI can significantly improve both speed and accuracy in payment processes, reducing the margin for manual errors. By capturing and analysing customer data in real-time, AI and automation minimises the need for time-consuming manual analysis, enhancing visibility across the entire payment lifecycle. This increased visibility also enables AI systems to boost payment security by identifying patterns indicative of fraud that human staff might otherwise overlook.
Moreover, AI and automation can help organisations analyse payment behaviour, supplier preferences and usage patterns to provide actionable insights. By anticipating needs and streamlining workflows – such as auto-suggesting preferred payment methods, optimising approval routes, or flagging anomalies in real time – AI and automation replicates the seamless intuitive experience found in B2C transactions. This results
in faster payment cycles, better financial planning and a more user-friendly interface for all parties involved.
While it has many benefits, integrating AI and intelligent automation into payment processes isn’t without its hurdles. Firstly, merging AI technologies within existing payment infrastructures can be complex and costly. There may be compatibility issues that require organisations to invest both time and resources to guarantee a seamless shift. In tandem, training staff to work alongside AI can present a challenge – especially without the correct training to interpret AI-driven insights effectively.
To truly unlock the value of AI in B2B payments, organisations must go beyond adopting new tools – they need to embed them in the right way. This means investing in integration, training and strategic alignment to ensure AI and automation enhances processes rather than adding complexity.
Bridging the B2B and B2C divide
The convergence of B2B and B2C payment models is no longer a trend – it’s a necessity. Organisations that continue to rely on legacy systems will struggle to compete in a market that values speed, transparency and user experience.
Forward-thinking businesses are already shifting towards payment ecosystems that resemble the consumer world – complete with mobile functionality, instant processing and AI-enabled insights. Those that lead this charge will set new standards for operational efficiency and customer satisfaction.
Ultimately, closing the gap between B2B and B2C payments means leaving behind the cheque books and embracing a smarter, faster, more connected financial future.
BEYOND PASSWORDS: BUILDING CONTINUOUS TRUST IN THE AGE OF AI-DRIVEN FRAUD
Identity theft has transformed from a lowtech nuisance—picture fraudsters rifling through discarded paper bank statements or stealing mail—into a hyper-industrialized, AI-driven enterprise that threatens the foundations of digital trust. Today’s attackers can order highly realistic driver’s licenses and passports online for under US$20, stitch together fragments of real data and fabricated details into synthetic personas that quietly build credit, and even deploy deepfake video and voice “liveness” spoofs to fool biometric systems. Meanwhile, the huge troves of personal data that companies collect and store are prime targets for hackers, potentially exposing millions of consumers’ identities in a single breach.
From Fake IDs to Synthetic Identities
What once required specialized equipment and expert forgers is now available as a selfserve online commodity. For less than the price of a pizza, a fraudster can download an AI-powered tool or visit a website that churns out high-quality IDs in minutes – complete with genuine-looking holograms, wear patterns, and printed microtext that evade common authenticity checks. In one widely reported experiment, a journalist purchased and generated a bespoke ID in under twenty minutes and successfully tricked a financial institution’s onboarding system.
But forging identities is only the start. Criminals are increasingly blending
snippets of real personal data such as birthdates, partial Social Security numbers, address fragments with wholly invented details to craft synthetic identities. These hybrid personas go through traditional Know Your Customer (KYC) pipelines, open line-of-credit accounts that look pristine, and then lie in wait. When they withdraw or default, the institution chalks it up to “bad debt” and often never traces the crime back to a human, making cleanup and restitution nearly impossible.
Compromised Verification Services and the PII Problem
Trusting a single external service to hold and verify your customers’ most sensitive personal data creates a glaring vulnerability. In a recent incident, a prominent identity-verification provider which is trusted by social media platforms, ride-share companies, and financial apps, exposed its administrative credentials online for over a year. This lapse potentially unlocked access to millions of uploaded selfies, scanned driver’s licenses, and other personally identifiable information (PII). Once attackers have that raw material, they can churn out realistic deepfakes, assemble new synthetic identities, or conduct largescale phishing campaigns against real customers.
Even beyond credential leaks, the sheer volume of PII stored in centralized databases magnifies the impact of any
breach. Every time an institution runs a KYC check, it sends more data into these vulnerable silos creating an ever-growing “honeypot” for cybercriminals. The old model where you collect data once, store it forever, and share it with every verification partner has become untenable.
Toward Consumer-Controlled PII
A more resilient approach moves the control of personal data from companies back into the hands of individuals. Imagine a system where:
• Individuals hold encrypted, cryptographically verifiable identity credentials on their own devices (or in a secure digital “wallet”), rather than uploading raw scans to a remote server.
• Verification requests trigger zeroknowledge proofs, allowing a customer to prove attributes (age, citizenship, license validity) without revealing the underlying document or raw data.
• Access permissions are granular and revocable, so a user can grant a lender the right to verify their identity this one time, then withdraw that permission immediately after—eliminating persistent exposure of personal images or numbers.
By embracing self-sovereign identity and privacy-preserving cryptographic protocols, institutions can dramatically reduce their centralized PII footprint, turn data breaches
into empty threats, and strengthen trust by giving users transparent control over who sees what, and when.
The Human Element: Social Engineering
Even the most advanced technical safeguards can be undone by a single phone call or email. In a May 2025 incident, cybercriminals impersonated employees of major UK retailers, contacting IT helpdesks with plausible backstories and urgent requests to reset privileged passwords. Without a robust process for validating those requests, helpdesk agents handed over the keys to entire internal networks without any forged ID or deepfake required. These breaches highlight that fraud prevention cannot be outsourced to technology alone. We need:
• Strict, documented escalation protocols for any request that could materially affect account security whether it originates from a customer, an internal user, or a third party.
• Ongoing social-engineering training and real-time risk scoring for support staff, so they recognize red flags and confirm identity through multiple independent channels.
• Solid linkages between real-world identities and user accounts, ensuring that when legitimate users need password resets helpdesk teams have a solid basis to determine if the ask is legitimate.
Daniel Flowe, Head of Digital Identity, LSEG Risk Intelligence
Daniel Flowe is Head of Digital Identity at LSEG Risk Intelligence, leading innovation in secure, interoperable digital identity solutions worldwide.
By reinforcing our “human firewall,” we ensure that even clever social-engineering attempts fail, protecting both institution and customer.
Building Digital Trust
To outpace AI-enabled fraud, financial institutions must adopt a continuous, adaptive trust framework grounded in three pillars:
1. Strong, ongoing linkage between signons and real world identities
Every access event, customer or employee, should be tied back to a verified identity in real time. Moving beyond point-in-time KYC, each login reinforces and refreshes the user’s identity context, preventing credentials from being reused or hijacked.
2. Leverage ‘good’ AI in the fight against ‘nefarious’ AI
Combine passive liveness detection, behavioral biometrics, device profiling, and network analytics into a single system of insight and intelligence. By using the same advanced algorithms that power deepfake creation, institutions can detect subtle anomalies, like mismatched typing rhythms or impossible geographic shifts, that manual checks would miss.
3. Secure, reusable digital identities
Issue cryptographically signed, portable credentials that customers control in a personal digital wallet. These reusable IDs minimize repeated exposure of raw PII, streamline onboarding, and empower customers to grant – and revoke – access on their terms.
Conclusion
Identity fraud has entered its most dangerous chapter yet. Forged documents, synthetic personas, and compromised verification services have rendered static checks obsolete, while social engineering continues to undermine even the strongest tech. Financial institutions have a choice: cling to outdated, vulnerable processes, or embrace a new paradigm of continuous trust – one that ties every access back to real identities, uses AI as a defensive tool, and returns control of personal data to the people it belongs to. By doing so, they won’t just reduce risk; they’ll redefine what it means to trust and be trusted online.
It’s no secret that traditional banking outreach methods are struggling to meet the demands of today’s digital consumer. Tactics like direct mail and word-of-mouth are no longer sufficient in a landscape where digital relevance is as important as community presence. Without a strong digital connection, even the most member-focused credit unions risk being overshadowed by national banks, fintechs, and large technology platforms.
Even paid digital advertising is becoming less effective. As of early 2025, the average cost to run a finance-related Google ad had skyrocketed to $313 per thousand impressions (CPM) – a sharp increase from just $3.30 in 2023. That’s a 9,400% increase in two years, with no guaranteed engagement – just visibility. Google’s rising ad prices reflect a shift in consumer behavior: people are searching less and instead relying more on curated, in-context experiences.
At the same time, consumer expectations continue to evolve. People no longer want to search for financial services; they expect these services to appear precisely when and where they need them. This change is fueling the rise of embedded finance. Instead of asking consumers to visit a branch, download an app, or search for the right product, credit unions can surface pre-approved, actionable offers within the digital environments consumers already use.
For example, thanks to embedded finance, a consumer browsing a digital car marketplace will encounter a real, pre-approved auto loan offer from a local credit union directly at the point of sale, without ever needing to search for financing
Barry Kirby, CRO and co-founder, Union Credit
HOW EMBEDDED FINANCE IS OUTPERFORMING PAID DIGITAL ADS IN CREDIT UNION GROWTH STRATEGIES
elsewhere. That level of contextual relevance not only increases engagement but also reduces friction and improves the likelihood of conversion. It also allows institutions to deploy resources more efficiently by focusing on targeted, high-intent interactions instead of costly mass-market advertising.
Embedded finance also expands the role of the traditional banking app. While standalone apps will remain important for managing accounts and financial planning, core actions such as applying for loans or opening new accounts will happen more naturally, directly within consumers’ daily shopping and financial experiences. In that sense, embedded finance doesn't replace a credit union’s app; it extends the institution’s reach beyond the branch physical and digital walls.
However, while the benefits of embedded finance are clear, the transition to embedded finance requires a major shift in both strategy and execution. Technology readiness is a major hurdle. Embedded finance depends on delivering real-time, actionable offers based on actual preapproval, not vague pre-qualification. Credit unions must be able to securely share underwriting models and product offers with third-party platforms while maintaining strong controls around data privacy and compliance. Without flexible, API-driven infrastructure, it becomes difficult to participate meaningfully in embedded ecosystems.
Equally important is the cultural shift. Credit unions have historically invested in drawing members into their own environments – branches, websites, and apps. Embedded finance turns that model
inside out, bringing the credit union to the consumer instead of the other way around. It’s no longer about controlling the full experience but about showing up in the right part of someone else’s.
Selecting the right digital ecosystems for embedded finance partnerships also matters. Not every platform aligns with a credit union’s mission or member expectations. Choosing trusted, community-aligned platforms and online marketplaces helps preserve the integrity of the credit union’s brand while extending its reach. Investments in technology will need to match this ambition. Modern, secure, API-ready systems capable of powering real-time offers will become foundational. Without the right infrastructure, even the best embedded finance strategy will struggle to deliver results.
The future of financial services won’t be defined by who has the biggest marketing budget or the flashiest mobile app, but by who shows up at the right moment, with the right solution, and in a way that builds trust. Credit unions should focus on understanding where potential members spend their time online, recognizing key life events that signal financial needs, and delivering real-time, personalized offers in those moments. Those that embrace this shift will extend their trusted relationships far beyond their own channels and into the daily digital lives of their members and prospects.
PRECISION VS. SCALE, OR WHY SMALL LANGUAGE MODELS ARE CRITICAL FOR HIGH-STAKES APPLICATIONS
Much like chemotherapy in medicine, highly targeted and designed to address specific conditions, SLMs excel in high precision-based settings. Whilst LLMs resemble broad-spectrum antibiotics: powerful, versatile, but often by far insufficiently precise for the most risksensitive applications. What is the best approach when it comes to regulatory compliance?
Artificial intelligence has become an indispensable tool for financial institutions to navigate regulatory complexities. Amid regulatory uncertainty and growing scrutiny, AI is playing an increasingly vital role in supporting compliance teams – by processing vast amounts of regulatory data, monitoring transactions and automating routine but essential compliance tasks.
For instance, consider an investment bank operating across multiple jurisdictions, each with its own evolving regulatory framework. Expert AI is able to track these regulatory changes in real time, reducing manual workload and allowing compliance officers to focus on strategic decision-making. However, not all AI models are equally suited for high-stakes applications. Just as different medical treatments target specific conditions, AI models must be tailored to their intended functions. The critical distinction lies between broad, general-purpose models and specialised, domain-focused AI solutions.
Large Language Models: versatility with limitations
Large language models (LLMs), such as GPT- 4, have become emblematic of AI's potential, offering automation across a wide range of tasks, from customer service to content generation. Trained on diverse datasets, LLMs excel at, for instance, summarising news articles, drafting content such as emails, or even writing code.
However, their broad training can lead to a lack of precision for specific enterprise applications. In high-stakes environments like regulatory compliance, medicine, legal analysis, where accuracy is of paramount significance,
LLMs can fall short. Their generalised nature increases the risk of producing misleading or incorrect outputs, a phenomenon known as "hallucination." McKinsey reports that in 2024, 61% of executives using generative AI tools cite LLMs as integral to their digital transformation strategies, but over 40% express concerns about accuracy and hallucination risks.
Moreover, bias in AI models presents another critical challenge. A recent FCA pilot study on bias in natural language processing (NLP) shows the risks firms face when deploying LLMs without accounting for bias. The study found that unchecked biases in AI-generated outputs can skew decision-making, misinterpret regulatory intent, and introduce compliance
risks. Ensuring fairness and accuracy in regulatory AI models requires a rigorous approach to dataset selection and ongoing monitoring of AI-generated content.
The substantial computational demands make them costly and sometimes impractical for specialised applications. A recent study from Bloomberg found that training and deploying LLMs at scale costs enterprises an average of $6 million per year, leading firms to seek more efficient, targeted models for high-precision applications.
The Rise of Small Language Models
In contrast, Small Language Models (SLMs) prioritise accuracy and efficiency over scale, making them particularly advantageous for specialised tasks requiring domain-specific insights. Trained on focused datasets tailored to specific enterprise needs, SLMs minimise inaccuracies and reduce the risk of generating irrelevant or misleading information.
A compelling example is the application of SLMs in regulatory compliance. By training on regulatory frameworks such as, for instance, MiFID II, Basel III, GDPR or any relevant other, SLMs ensure compliance officers receive accurate and actionable insights rather than generic AI-generated responses.
At Corlytics, our strategy is to develop specialised SLMs trained on carefully
curated, structured regulatory content to enhance accuracy in compliance workflows. These models can parse complex legal documents, track jurisdiction-specific rule changes and flag potential compliance gaps with greater accuracy than their LLM counterparts.
By focusing on precision, SLMs address the shortcomings of LLMs, including bias, hallucination and inefficient resource consumption. Their smaller model size also ensures a significantly lower carbon footprint, making them a more sustainable and cost-effective option. Additionally, and very importantly, customer data remains securely within the client’s virtual infrastructure versus being exposed to 3rd-party model providers, and this ensures compliance with strict data privacy regulations.
SLMs vs. LLMs: AI to Specific Enterprise Needs
While LLMs provide versatility across a broad range of applications, their generalist approach limits their effectiveness in fields requiring high accuracy. SLMs, by contrast, leverage targeted training and efficient resource utilisation to deliver outputs that are both reliable and contextually relevant.
For financial compliance teams, the ability of SLMs to process and interpret regulatory documents with precision helps institutions remain compliant with
John Byrne, Founder and CEO, Corlytics
John Byrne is the founder and CEO of Corlytics, leading the company’s vision, strategy, and global growth in regulatory intelligence and compliance automation. A seasoned fintech entrepreneur, he previously founded Information Mosaic into a global leader in securities software, acquired by Markit in 2015. Since launching Corlytics in 2013, he has earned multiple RegTech awards and made Corlytics the first in its field to achieve ISO 42001 certification for responsible AI. A Stanford Business School graduate with a background in electronic engineering, John also serves on the board of CeADAR and is a regular speaker at global events like SIBOS and RegTech Summits.
evolving regulations, thereby reducing the risk of fines and reputational damage. By focusing on high-quality, domain-specific data, SLMs enable more informed decisionmaking, reinforcing their value in critical applications where precision outweighs scale.
Much like chemotherapy in medicine, highly targeted and designed to address specific conditions, SLMs excel in precision-based settings. In comparison, LLMs resemble broad-spectrum antibiotics: powerful, versatile, but often insufficiently precise for the most sensitive and high-stakes applications. In regulatory compliance and other fields where accuracy is non-negotiable, SLMs emerge as the superior choice, offering tailored AI solutions that align with the rigorous demands of specialised industries.
A hybrid Approach
For compliance teams, the ability of SLMs to process and interpret regulatory documents with precision helps institutions remain compliant with evolving regulations, thereby reducing the risk of fines and reputational damage. By focusing on high-quality, domain-specific data, SLMs enable more informed decisionmaking, reinforcing their value in critical applications where precision outweighs scale.
However, the most effective AI-driven compliance strategies do not rely solely
on one type of model. Instead, companies like ours for example, are increasingly combining LLMs, SLMs as well as other machine learning techniques to create hybrid AI solutions that are capable of balancing scale with high accuracy.
LLMs, with their extensive linguistic capabilities, can parse vast amounts of general regulatory text and extract relevant information. Whilst SLMs ensure that outputs meet the precise standards required in regulatory environments.
For example, a financial institution tracking real-time changes in global sanctions lists might use an LLM to scan and summarise regulatory updates, while an SLM refines these summaries into actionable insights tailored to internal policies. Similarly, a firm conducting transaction monitoring could use bespoke machine learning classification models to flag potential risks before an SLM assesses the flagged data against jurisdictionspecific compliance rules. This layered approach ensures that AI-driven solutions are not just powerful but also extremely precise, cost-effective and secure. The application of small language models in compliance is indeed proving that less is more.
There’s a striking irony at the heart of today’s banking industry.
While consumer-facing innovations like real-time payments project a cuttingedge, tech-savvy image, the reality behind the scenes tells a different story. Many banks still depend on outdated legacy IT infrastructures to support both their customer-facing products and internal financial operations. Financial closes that stretch beyond 10+ days are still the norm, and ageing items often cause further delays as spreadsheets are shuffled between teams, email threads multiply, and manual reconciliations drag on.
Today, being high-tech and global is not a luxury for banks — it’s a baseline expectation. So, the gap between polished digital products and ageing internal systems isn’t just ironic; it’s dangerous. Legacy technology opens the door to costly errors, exposes institutions to cyber risks, and forces finance teams to waste valuable time on manual tasks that should be automated.
The Hidden Costs of Legacy Systems in Banking
According to a report by the UK Treasury Committee, a sub-section of the major UK banks and building societies had at least 803 hours of unplanned tech and systems outages in the two years leading up to March 2025. This equates to more than 33 days of outages impacting both customers and internal operations. These outages were caused by various factors, with some saying outdated legacy technology systems were part of the problem. Additional factors contributing to the outages included issues
Darren Heffernan, CEO, Trintech
BANKS ARE ALL IN ON REAL-TIME PAYMENTS
– SO WHY ARE THEIR OWN FINANCIAL SYSTEMS STILL STUCK IN THE PAST?
with third-party suppliers, disruption caused by system changes and internal software malfunctions.
Additionally, companies operating in banking and payments have specific accounting requirements. These include complex financial instruments, loan provisions, fair value measurements, and hedge accounting. This can make embedding new technology a tricky job, meaning legacy systems are not regularly upgraded. When integrating automation into their financial systems, banks must navigate strict regulatory requirements while ensuring they can manage the massive volumes of data inherent to the industry.
Bridging the Gap: A Strategic Path Forward
Modernising financial systems may be a daunting task, but it’s a critical step for banks looking to stay competitive and resilient.
As banks rethink their finance and accounting infrastructure, they must evaluate technology partners not just on features, but on industry-specific expertise, regulatory awareness, and the ability to integrate within a broader ecosystem. Integration is often the Achilles’ heel: when systems can’t communicate, data silos form, manual workarounds proliferate, and errors creep in.
Consider the financial close process – a complex, data-heavy workflow requiring information from disparate, outdated systems. Without proper integration, transaction matching becomes a manual burden, delaying the process by weeks. In
some cases, month-end close can stretch beyond 30 days, creating deadline pressure and stress for finance teams.
However, by investing in technologies that support interoperability, banks can dramatically reduce manual intervention and accelerate workflows. For example, modern financial close automation solutions can cut close times by more than 50%, freeing up resources for more strategic initiatives and enabling real-time insights.
Conclusion: True Digital Transformation Starts Within
Upgrading financial infrastructure is no longer optional – it’s foundational. While external innovation grabs headlines, banks must also look inward. Modernising core finance systems not only reduces operational risk but also unlocks the full promise of digital transformation. It’s time banks brought the same energy to their internal processes as they do to their customer-facing platforms.
Darren Heffernan is the CEO at financial close and reconciliation software provider Trintech. He joined Trintech in 2001 and has held a number of strategic and corporate development roles during his tenure, including CFO and, most recently, President & COO. Driving numerous acquisitions and expansions into new markets, Darren has been instrumental in leading Trintech’s growth strategy.
The financial services industry stands at a pivotal moment as generative AI (GenAI) promises to re-shape treasury and finance functions. With over 84%* of treasury professionals anticipating significant AIdriven change within the next two years, the imperative is clear: organisations must move beyond automation and embrace intelligent decision-making, security, and strategic agility to thrive in a rapidly evolving landscape.
From Back-Office to Strategic Partner
Modern treasury teams have become core drivers of financial resilience and agility. With 78% of professionals ranking their influence on the bottom line as a top priority, and over 54.5% highlighting fraud detection as a significant challenge, it’s no surprise that organisations are accelerating their adoption of GenAI solutions. However, unlocking GenAI’s full potential requires more than advanced algorithms; it demands a foundation of trust, transparency, and adaptability.
Enabling Trusted GenAI Adoption: The Pillars of Success
As organisations consider how to safely and effectively adopt GenAI in finance, several foundational elements have emerged as critical for success:
• Security, Privacy & Compliance by Design
Given the sensitivity of financial data, any GenAI deployment must prioritise robust data security, privacy, and regulatory compliance. The most effective models are those hosted in controlled environments, with strict permissioning and auditability— ensuring that data never leaves the organisation’s secure perimeter and that
Morné Rossouw, Chief AI Officer, Kyriba
THE ESSENTIAL ELEMENTS FOR TRUSTED TREASURY TRANSFORMATION
every AI-driven action can be traced and validated.
• AI Powered by Unified Platform
Fragmented, bolt-on AI “point solutions” create operational silos and risk. Industry leaders are converging on unified, platformbased AI strategies that embed agentic intelligence throughout the treasury ecosystem from cash forecasting to fraud detection and working capital optimisation. This approach delivers seamless user experience and faster, more reliable outcomes.
• Conversational, No-Code Interfaces
To democratise the power of GenAI, next-generation treasury platforms are moving away from technical, code-heavy interfaces. Instead, they leverage intuitive, conversational user experiences allowing finance professionals to interact with AI in natural language and receive actionable insights without specialist training. This is crucial for bridging the digital skills gap and accelerating adoption.
• Transparent, Explainable AI
Trust is the bedrock of GenAI adoption. Open-source reasoning models and transparent AI “thinking” processes allow finance teams and auditors to understand, verify, and explain how decisions are made. This transparency is essential for building stakeholder confidence and meeting both internal and regulatory demands for explainability.
• Co-Innovation and Continuous Feedback
The most effective AI solutions are developed in close partnership with end users. Leading organisations employ feedback loops, focus groups, and design thinking workshops to ensure that AI
capabilities address real-world pain points and evolve with changing business needs. This co-innovation model ensures GenAI delivers measurable business impact, not just theoretical promise.
• Extensibility and Customisation
Each finance organisation faces unique challenges. GenAI platforms that enable users to build, tailor, and “teach” their own intelligent agents within a secure, governed environment unlock agility and innovation at scale. This flexibility is a key differentiator in a landscape where one-size-fits-all solutions often fall short.
The Road Ahead
As the sector races to realise the benefits of GenAI, key performance indicators—such as faster onboarding, reduced manual effort, improved cash conversion cycles, and the automation of up to 35% of business tasks— are becoming the new standard. However, the true competitive advantage will belong to those who blend technological innovation with a foundation of trust, transparency, and collaboration. By adopting these pillars, finance organisations can transform treasury from a reactive, back-office function into a proactive, strategic partner—ready to navigate volatility, seize opportunity, and lead the next wave of financial innovation.
* Source: IDC InfoBrief, May 2025, "Agentic AI in Treasury: Navigating Trust and Realizing Potential"
USING REAL-TIME DATA TO ENHANCE DECISION-MAKING
Mat Gapp, Head of Vision Next Product, Fiserv
As the now familiar saying goes, ‘data is the new oil’. In today’s digitalised world, data has emerged as arguably the most valuable asset, shaping decisions, driving innovation, and redefining how organisations operate.
Like oil, data's true value lies in how it’s extracted, processed, and used. With today’s economy becoming increasingly digitalised, data has become an incredibly lucrative industry. It is now being utilised across industries to enable valuable product insight that can drive revenue for businesses, in turn enhancing strategy and improving customer experience.
A lot of financial institutions have increasingly developed their business models around managing extensive datasets from a variety of origins. These businesses leverage vast volumes of data to create business and consumer products that address customer issues and meet market demands.
Putting customers at the heart of strategy
For financial institutions, the key element of fine-tuning their strategy, and, ultimately, a major contributor to both their short- and long-term success, is ensuring customers are at the centre of their approach. Ongoing engagement with customers is vital in understanding trends in consumer behaviour, as well as customer needs. Only through gaining these insights can financial services firms offer their customers products and services that deliver real value to them, and genuinely meet their requirements.
By prioritising the customer and being acutely aware of their needs, financial institutions build client loyalty. If a bank or insurance provider fails to personalise their approach, they run the risk of losing clients to more adaptable competitors. Maintaining customer loyalty is critical – to ensure retention rates are high, financial institutions must keep clients at the heart of their strategy.
Data-driven decision-making is at the centre of this message. Financial organisations can provide greatly improved services to their customers through utilising in-depth data to analyse and understand customer behaviour in real time. By incorporating this data into the decisionmaking process, financial institutions can tailor their offerings to specific events or changes in customer behaviour, ensuring
that what they’re presenting to clients and consumers is much more relevant.
Technology
at
the forefront of data-driven decision making
Another major development is the use of platforms that incorporate data cloud and artificial intelligence (AI) technologies to create a single central source of data. These offerings are accompanied by full lineage and change history that is readily accessible and available for use by multiple applications and products. As a result, complex data pipelines are not required – instead, data is processed in a far more streamlined and efficient way.
In the financial services industry, card management platforms are at the forefront of data-driven decision making. By utilising features like real-time data, APIs and cloud scalability, cardholders can control and customise their card preferences. This leads to a much more seamless and personalised experience for customers when managing their cards.
Utilising data to benefit issuers and their customers
For both financial institutions and their customers, card management solutions provide ample benefits. From the point of view of consumers, these tools provide an easy-to-navigate interface, creating a simple, accessible experience for users. These offerings streamline processes for cardholders – cards associated with various accounts, ranging from debit accounts to loan accounts and all others, facilitate smoother transactions for customers.
As well as this, card management platforms allow customers to take control of their finances and make informed decisions regarding their money. Real-time data allows cardholders to remain up to date on their card management options and gives users complete transaction visibility. Cardholders can also set personalised spending limits based on their own budgets, informed by real-time data analytics.
For issuers, the data offerings that card management platforms present gives them the edge over competitors. Access to real-time data on customer behavioural patterns and live customer feedback allows issuers to proactively modify their strategies and enhance operational
efficiency, without having to depend on traditional batch processing.
Personalisation also allows issuers to stay competitive. Through data analytics offered by card management platforms, card issuers can offer their customers uniquely tailored services. This includes marketing efforts, loyalty programmes, discounts and incentives. Additionally, issuers can use their access to live data to move early and keep ahead of the curve with their marketing.
Card management solutions also strengthen financial safety for both customers and issuers by preventing fraud. For instance, the disposable card feature for customers streamlines virtual transactions and minimises the risk of financial fraud. This is done by using unique, one-time card details that are automatically unauthorised after a transaction is completed. Users receive new details for every transaction, so their information cannot be guessed or stolen. For issuers, fraud protection services deliver robust security measures, as well as real-time monitoring and data analytics.
The use of data in elevating decision making and enhancing financial offerings is growing in prevalence. Financial institutions that ignore the opportunity to follow suit risk losing customers to issuers who are taking a more customer-centric approach. By utilising card management platforms, issuers not only have instant access to the latest data, enabling them to keep their finger on the pulse, but they can also provide consumers with a scalable and fully customisable approach to financial management.
With over 25 years of experience in the Cards and Payments industry, Mat Gapp has dedicated his career almost entirely to the Vision family of products. He began his journey as a COBOL developer in 1999 and has since embraced a variety of roles including Production Support, Project Development, Architecture & Design, Business Analysis, Client Consultancy, and Delivery. Mat now leads the creation of the new Vision Next platform from a Product Management perspective.
COMMERCIAL BANKS SUPERCHARGE FINANCIAL AND ERP PLATFORMS WITH EMBEDDED FINANCE
Commercial banks are getting wise to the immense potential of embedding financial services directly into existing business systems, chiefly ERP and financial platforms.
This gathering force is now changing how businesses manage finances, initiate payments, and access banking services – all without leaving their preferred collaboration platforms.
In the process, embedded finance reshapes how commercial banks and their customers manage business payments while gaining access to new prospects.
Embedded Finance = New Channels for Banks
The adoption of embedded finance continues to accelerate due to compelling benefits on both the revenue and cost fronts. And the ease of managing business payments from inside familiar ERPs and finance systems drives the demand for integrations, from small businesses to enterprises.
Embedded finance effectively gives banks a new channel to market. You can go after a great many small and mid-sized enterprises (SMEs) by generating an integration with just one widely used finance package.
And SMEs are only one group benefiting from a partnership approach that allows banks to expand their reach without using outdated customer acquisition methods. By integrating with popular upstream
platforms, banks can instantly access millions of potential customers who already use these systems daily.
Bottomline itself is taking embedded payments to banks and financial institutions worldwide, and finding enthusiasm usually reserved for AI (more on that in a moment).
Regarding costs, embedded finance delivers significant operational efficiencies through tighter straight-through processing (STP). Traditional methods require too much manual intervention, like exporting payment files from ERP systems and uploading them to banking portals. This outdated process makes payments too timeintensive while pushing up costs.
Embedded finance is known to eliminate costly extra steps, reduce errors, minimize fraud opportunities, and accelerate transaction times. Large banks are particularly invested in embedded finance as both a revenue source and an attractive option for clients.
As would be expected, the biggest banks are all over this trend. They're investing in embedded finance because they see it as a revenue driver, and as a way of demonstrating their operational sophistication to prospects and customers.
Overleaping Tech and Regulatory Challenges
Despite its many benefits, implementing embedded finance often presents
challenges for banks. A primary hurdle right now is ensuring resilience in API infrastructures.
The biggest obstacle here is robustness and reliability. If you're offering an API that's going to be called 24/7, you need to make sure that it's up 24/7.
Banks must also navigate complex regulatory requirements while maintaining security. Every transaction initiated through embedded finance still requires the same regulatory and KYC checks as traditional banking channels.
Some banks have invested substantially in this area, developing sophisticated systems to perform these checks in real-time without disrupting the user experience.
Increased connectivity also brings new cybersecurity concerns. Cyberfraud risk management is essential as more B2B payments systems transact with more accounts.
There's an increased risk of cybercrime when opening your systems to more interactions with more integrations. Banks need stout security measures to protect against these threats while maintaining seamless service delivery.
Additionally, banks need operational maturity to provision and support failsafe integrations. This maturity calls for significant investment in infrastructure and personnel to ensure systems always remain available and responsive. Banks that have already invested in becoming
Sadiq Javeri, Head of Platform Services – Financial Messaging, Bottomline
Sadiq Javeri is responsible for Bottomline’s go-to-market enablement and product partnerships. With over 18 years of experience, Sadiq is a seasoned speaker and regular contributor to key industry media and events, sharing insights on emerging trends and thought leadership. Before joining Bottomline, Sadiq honed his expertise in working for and in corporate banks, including his tenure in Investment Banking at Deutsche Bank and as a Financial Services Consulting at Oliver Wyman. He has successfully defined, developed, and commercialized fintech products, leveraging his deep domain knowledge in the financial messaging landscape, payments modernisation as well as treasury and capital markets.
Sadiq holds a BEng in Information Systems Engineering from Imperial College London and an MPhil in Finance from the University of Cambridge.
API-ready (or partnered to accomplish it) enjoy distinct advantages.
Modernizing Operations and Cash Flow
Embedded finance is evolving how businesses manage their financial operations, with payment processing and cash flow as prime beneficiaries. By integrating financial services directly into operational systems, banks and businesses are now achieving unprecedented levels of automation, throughput, and efficiency.
Once a company can connect to a bank via an API, you don't have to manually tell a website what financial transaction you want to perform. You can put code around it, saying, “If this payment comes through from this vendor, wait three days, then send it."
Automation like this can extend to payment scheduling, reconciliation, and cash management. Businesses can set rules for different payment types, vendors, or currencies, allowing for more strategic cash flow management. Real-time visibility into financial positions enables better decision-making and resource allocation.
For mid-sized companies especially, embedded finance is an important step. Organizations in this tier often struggle with the move from manual processes to automated systems. Embedded finance provides a bridge, allowing them to automate certain operations without
investing in complex, standalone financial systems.
Benefits include:
1. Reduced manual intervention and associated errors
2. Faster payment processing and improved cash flow visibility
3. Automated reconciliation between operational and financial systems
4. Enhanced security through the elimination of manual file transfers
5. Better vendor relationships through more predictable payment timing
AI and New Embedded Possibilities
The embedded finance terrain shows several emerging trends poised to model its future. My own view is widely held: AI is playing a progressively important role in financial decision-making and business payments. Over time, this will cause more change.
Once you can put rules and code around how you want to manage your liquidity, you can put AI around it, too. The ability to tell an AI to take data and do what makes sense in the context of your cash management needs is a notable development.
Early adopters of embedded finance are now looking beyond basic payment functionality to more advanced use cases. These include embedded lending, insurance, and more complex financial services integrated directly into platforms.
Market watchers are comparing this year’s series of financial system upgrade deadlines to previous technological shifts. This time around, nothing has led to true obsolescence. The banking portal will remain relevant for complex transactions and regulatory oversight.
In the near term, we will likely see banks partnering more extensively with fintechs and payment service providers (PSPs) to deliver embedded finance solutions. Banks will need to balance maintaining control over complex, high-value transactions, while allowing partners to handle more routine financial operations.
As the ISO 20022 messaging standard becomes more widely adopted, the potential for embedded finance will expand dramatically. This standardization will enable more seamless data exchange between systems, supporting more advanced financial services embedded within business platforms.
For banks and corporates alike, embedded finance represents not just a technological shift but a fundamental reimagining of how B2B financial services are delivered and consumed. Those who welcome this trend stand to gain serious competitive advantages in efficiency, customer experience, and market reach.
CX IS THE NEW BATTLEGROUND FOR BANKS— AND INSTANT PAYMENTS ARE A FRONTLINE ADVANTAGE
Todd Clyde, CEO, Token.io
Europe’s banking industry is at a critical juncture.
In recent years, banks have felt the impact of economic headwinds and market volatility, which have constrained opportunities for growth. While this outlook is now improving, a new challenge has emerged: customer experience (CX) is in decline.
According to Forrester’s 2024 European Banking Customer Experience Index, CX quality dropped significantly last year.1 This trend is concerning. In today’s competitive market, CX is a critical driver of customer loyalty and long-term revenue growth.
There is, however, a significant opportunity on the horizon. The Instant Payments Regulation (IPR) is a chance for banks to not only elevate their CX — but to also take the lead in payments innovation.
What does instant payments regulation mean for banks?
In Europe, instant payments — characterised by speed, convenience, and 24/7/365 availability — are becoming the new standard. The SEPA Instant Credit Transfer (SCT Inst) scheme, available since 2017, is now being mandated.
While SCT Inst is already supported by many banks in the Single Euro Payments Area (SEPA), it is not supported by all. As a result, SCT Inst’s share of total SEPA (SCT Inst and CT) volumes was just 21% as of Q4 2024.2
Europe’s Instant Payments Regulation set out to address this. It is accelerating the rollout of instant payments and normalising them across Europe by mandating banks in the Eurozone to support SCT Inst for incoming and outgoing payments. Crucially, they must offer the service at the same or lower cost than standard credit transfers.
The parallels with PSD2 are striking
The IPR is being enforced in two key phases. The first regulatory milestone passed in January, meaning all banks in the Eurozone must now be able to receive SCT Inst payments. Looking ahead, banks are now sprinting to meet the October regulatory
deadline, which requires that they can send SCT Inst payments.
Given the implementation challenges, it’s no surprise that some banks are concerned about the regulation. It means upgrading legacy infrastructure, conquering operational issues and ramping up fraud prevention –and all at breakneck speed. As a result, 58% of European banks3 that do not currently offer instant payments believe that the implementation timelines are unrealistic.
Yet, the banking industry has navigated similar transitions before.
The enforcement of the Second Payment Services Directive (PSD2) in 2018 was initially met with apprehension by banks. Yet, ultimately, PSD2 has been a driving force for payment innovation — and continues to unlock new revenue streams and growth opportunities for banks. Forward-thinking banks leveraged open banking APIs to launch new account-toaccount (A2A) solutions, benefiting both corporate and retail clients.
The starting gun in a new race
The resulting impact of PSD2 on payments has been profound. Open banking-enabled A2A payments, often called ‘Pay by Bank,’ have successfully delivered a convenient, fast and secure digital payment method that is accessible to nearly anyone with a UK or Eurozone bank account. Instead of entering card details, customers simply pay directly from their bank account, easily and securely authenticating the transaction in their mobile app.
Although changing payment behaviour is a formidable challenge, the data speaks for itself. Pay by Bank is trusted, familiar and widely adopted. Analysts4 predict that by the end of this year, one in four Europeans will embrace it for online transactions, and three in four will use them by 2029. By 2030, European’s use of Pay by Bank for e-commerce is expected to eclipse all other digital payment methods, second only to wallets.5
The Instant Payments Regulation builds on this momentum, and will enable banks to further differentiate themselves through product and CX innovation.
4 Juniper Research, Global A2A Payments Market Research Report – Forecasts 2024-29
5 Worldpay, 2025 Global Payments Report
Taking CX to new heights
In this respect, the Instant Payment Regulations should not be viewed as a burden. They are the starting gun in a race that will separate innovators from laggards.
Instant payment capabilities can transform everyday banking and take CX to new heights for consumers by making real time deposits, bill payments, and loan repayments standard practice. This shift will enhance operational efficiency and reduce costs. Crucially, it will significantly improve customers’ experiences by making it easier to transfer money between accounts across different banks and by streamlining bill and credit card payments.
Leading institutions are already responding to the opportunities. HSBC, for example, has introduced ‘HSBC Open Payments’, combining the speed of card payments with the simplicity of bank transfers, helping its corporate clients reduce processing costs and improve cash flow while offering consumers a secure and convenient payment option.
Realising the vision of homegrown, pan-European payments
Ultimately, the Instant Payments Regulation represents more than a regulatory requirement. It is an opportunity to deepen customer relationships, drive revenue growth and gain a competitive edge.
It is also an opportunity to realise Europe’s vision of a homegrown European payments system. Amid rising geopolitical risks, the debate around Europe’s payments sovereignty is heating up, with the European Central Bank outlining the need for a “European payments offer.”
While different solutions are being explored, Europe does not need to reinvent the wheel. We’ve already driven significant growth in Pay by Bank. Today, nearly anyone with a bank account in any of the 27 EU member states can use it to easily and securely pay for goods and services.
The market is keen, as is the regulator. The time is right to deliver on the promise of an instant, seamless, and secure European payment system.
All that remains is for the banking industry to deliver.
Buchwaldt-Nissen, SVP, Head of Hospitality, Nexi Group
Simon
WHY PARTNERSHIPS AND PERSONALIZATION ARE THE FUTURE OF PAYMENTS
The future of payments will see solutions become increasingly embedded into industryspecific software, enabling businesses to streamline operations and leverage valueadded services, transforming how they operate. It will ultimately create a better experience for the customer and simplify management for the merchant.
By offering payments embedded within the systems of independent software vendors (ISVs) such as Adobe, Shopware or one of the more than 500 others we work with across Europe, payment providers can help businesses grow faster and accept almost any payment method instantly. These integrations will also enable the payment function to operate seamlessly alongside other business processes for a more convenient, simplified management experience.
For example, a restaurant could use one integrated commerce system to take bookings, allocate tables, assign waiting staff, manage inventory and provide trusted ‘invisible’ consumer payment options. Doing so can achieve significant operational efficiencies while providing a more consistent, highquality customer experience.
It can also enable a new approach focused on ‘hyper-personalization’. Based on integrated, shared data, individual preferences of guests could become seamlessly known and understood by a hotel, for example. Enabling guests to feel recognised at every touchpoint, from personalised emails to in-room experiences, will play a crucial role in creating a unique, personalised and memorable stay.
One System for Endless Personalization
The growing demand for personalized experiences is being driven in part by consumers that value the ability to pick and choose their preferred payment method, which can vary depending on the setting or scenario.
At a restaurant, for example, table-side payment systems and contactless options can
reduce wait times, while enabling customers to split bills easily, leaving a lasting positive impression. In retail, online and physical environments will continue to merge, delivering more seamless experiences: for example, a customer might browse items online, reserve them, and complete the purchase in-store.
In each of these environments, every interaction plays an important role in contributing to the customer’s overall satisfaction: a poor experience in-store, a clunky user journey online, or any complexity in choosing the goods/services and paying for them, threatens customer retention. It’s clear that payments play a pivotal role in shaping these experiences.
The Next Frontier for Payments
Payments are moving beyond transactional utility to deliver specialized, value-added payment experiences tailored to each vertical. Again looking at the hospitality industry as an example, we can see competition intensifying and businesses constantly looking for ways to enhance the customer experience. Digitalization is helping hotels rethink their offer, putting more emphasis on personalization and creating positive memories that last a lifetime. Conversely, any delays or complications in paying will also create memorable experiences, but for all the wrong reasons.
This highlights a critical point for the future of payments: to work effectively, it must be available alongside other tools within one single system, so that each business can manage its operations efficiently.
What this means is that fintechs can easily address the needs of different verticals with a simple ‘plug-and-play’ offer that ISVs and software platforms easily integrate into their merchant offer, leveraging the completeness, scalability, security and support of proven solutions. Integrated payment systems will ensure consistency across channels, no matter
which vertical the merchant is in, with no discrepancies in pricing, discounts, or loyalty rewards.
Some might refer to this growing trend as “Payments Inside” – easily-integrated, scalable, high-performance payments software, designed to support sector-specific requirements. To succeed, technology providers need to realise they are stronger together: partnerships will be the key to unlocking the potential of personalized payments for all. Collaborations, like the one Nexi has with Microsoft, will give thousands more merchants the opportunity to onboard quickly and offer fast, efficient, and secure card payments to customers, no matter how large or small the business.
Let’s face it, no merchant starts its journey by asking itself what payment provider it should use; it’s up to payment service providers to ensure merchants can easily unlock a comprehensive payment offer via seamless integrations and value-added services. This is why Nexi and Planet recently announced a partnership to combine omniacceptance for local schemes and alternative payment methods with vertical capabilities including acceptance technology, platforms integration and dynamic currency conversion.
Catering to diverse customer preferences is critical in the market. Businesses that offer a wide range of payment options and enhanced personalization during the payment process will not only meet expectations but exceed them. If fintechs do this job well, consumers in the future will barely notice us at all, putting the focus where it should be: effortlessly enjoyable experiences.
Zitah McMillan, CEO, Finexos
IT’S TIME TO TRANSFORM AFFORDABILITY ASSESSMENTS WITH AI
Affordability assessments are a fundamental part of responsible lending, serving to protect both consumers and lenders from the risks associated with unaffordable debt. Their importance has grown significantly in recent years, driven by regulatory requirements, economic volatility, and a renewed focus on good consumer outcomes.
Whilst there were no new affordability rules specifically required by the Consumer Duty Act, the FCA has reinforced the need for credit providers to have a robust and explainable approach to affordability as part of the overall creditworthiness assessment. The aim is to deliver good outcomes for retail consumers and avoid foreseeable harm or detriment.
The issues with outdated affordability assessments
Affordability assessments support the evaluation of whether a borrower can realistically manage repayments alongside their essential living costs, such as rent, utilities, groceries, and other financial commitments. This process helps ensure that taking on new debt will not push individuals into financial hardship or force them to sacrifice basic needs to meet loan obligations.
Yet traditional affordability assessments, rooted in static credit reports and fixed metrics, are increasingly inadequate in today’s dynamic economic landscape. These outdated methods can exclude viable borrowers, perpetuate systemic biases and fail to account for real-time financial behaviours. Lenders are naturally concerned about potential default risks and impact on the overall health of their loan book or portfolio, they also realise that they are missing out on potential business and revenue growth by using outdated
models. Balancing these competing pressures is vital for lenders.
When the time comes for the Board level annual assessment of whether their firm is delivering good consumer outcomes, which includes reviewing their approach to affordability, firms need to be confident that they can evidence and justify their decisions. Policies and procedures form part of this assessment, as does the information, methods and models used to make the lending decision.
The future of affordability assessments
Each borrower’s financial circumstances and how they manage their finances is unique. On the regulatory front, the FCA requires that every customer is treated as an individual and that decisions are made in a borrower’s best interests. There is also a renewed and welcome focus on financial inclusion, by the Government and Regulator, which is natural after a period of financial challenges for consumers.
Combining Artificial Intelligence (AI) with real-time transactional and behavioural data analysis has the ability to transform affordability assessments into fairer, more accurate tools for lenders and borrowers alike. This method delivers a holistic understanding of a borrower’s financial reality and if they can meet repayments without suffering financial hardship.
Individuals make different sacrifices and choices when managing their finances. Understanding discretionary spending is essential to achieving the most accurate affordability assessments as discretionary expenses are more flexible and provide borrowers with room to adjust their spending during periods of financial strain.
Analysing transactional data delivers
granular details of income and expenses, allowing for precise differentiation between fixed and discretionary spending. It also supports the identification of emergency situations and long-term financial pressures which are critical for ongoing monitoring of loan affordability. Financial health is fluid and without utilising more up-to-date or real-time data to assess affordability, a genuine assessment of affordability is simply not possible.
The time is now
It’s time to move beyond basic income, high-level outgoings and repayment tracking. Transactional and real-time data along with advanced AI analysis provides a far more accurate picture of a borrower’s ability to repay and, crucially, allows us to determine whether borrowers can maintain their quality of life while repaying loans.
It means that firms can tailor their approach to the needs of their target consumer and in particular if there are vulnerable consumers as part of that target.
It means Boards can see the evidence and justification of the lending decisions made across their portfolio and understand how the affordability assessments contribute to good consumer outcomes.
To ensure truly responsible, long-term lending decisions that deliver better access to credit and outcomes for consumers and more sustainable lending for credit providers, we should be looking to AI for the future of affordability, today.
Nathaniel Powell, CEO and founder, Deep MM
Nathaniel Powell is the CEO and founder of Deep MM, an AIpowered credit trading solution providing the most accurate U.S. corporate bond pricing to the private credit market.
DEEP LEARNING, LEMS AND THE FUTURE OF QUANTITATIVE FINANCE
Generative AI is changing the world rapidly with commonplace modalities like natural language, general text (such as programming languages), images, video, speech, and music. Perhaps the most important modality of all, however, is overlooked by investors and the vast majority of AI researchers: events.
Large language models work by predicting the probability of the next token (similar to a word) in a string of words. For example, when generating an essay the large language model predicts the probability of observing each of the tokens in its vocabulary and then samples from that probability distribution. Depending on the generation rules, it may emit the token with the highest probability or in some cases, it might be configured to randomly generate the tokens according to their predicted probabilities. For a typical Wall Street quantitative financial use case, this might help with tasks like information extraction (like news analysis), but it won’t help directly with predicting the probability of trade prices.
This begs the question, has quantitative finance been left behind by the generative AI revolution?
The Time Comes for Deep Learning
I recently had the privilege of speaking at Columbia Mathematics of Finance (MAFN) 2025 Future of Portfolio Management & Artificial Intelligence Conference where the audience and speakers were mostly made up of quantitative finance industry leaders. I found that when they spoke of deep learning, it was all about large language
models and their typical applications within banking and finance (e.g., Customer service, information extraction), with occasional references to deep learning as a future solution for quantitative finance scenarios. Anecdotally speaking, it seemed I was the only one who was strongly advocating for deep learning in quantitative finance today.
The General Purpose LEM
When a general purpose model is trained on a set of asset classes (like stocks, bonds, commodities, and FX), in theory it should be able to outperform single asset class models which were only trained on each of the asset classes individually. Neural networks of sufficient depth/complexity have the capability to learn more patterns about the underlying nature of financial markets with more training data. This is analogous to how a general purpose LLM is able to outperform on many tasks versus models which have only been trained on data from a single task. For instance, a state of the art general-purpose LLM, when only trained on poetry data, will not perform as well as the same LLM model trained on all of the internet’s text, including poetry but from many other domains as well. This is because many of the principles of language learnable from other domains, such as social media, Wikipedia, news, blogs, etc., are also applicable to poetry.
At Deep MM we have developed what we call a “Large Event Model” (LEM). It’s not predicting the next token of a sentence –it’s predicting the probabilities of attributes of the next financial event. Instead of having as context one document, it typically needs to be considering the whole market.
With US corporate bonds, our model is taking into account all of the related bonds as well as changes to macro indicators.
To create this general purpose LEM, we are inventing a new data format which will be self describing in the sense that it communicates the relationships between the input data and the requested output event probability. This coupled with a flexible deep learning architecture will allow the model to generalize to a variety of asset classes, so instead of manually repricing millions of securities, human traders will have more time to provide much more valuable services, such as negotiating big trades, building relationships, managing risk, creating and updating trade and investment strategies, etc.
In time these large event models can also be applied to non-financial event streams, like a medical event or the arrival of a truck at a warehouse.
LEMs For The Future
Beyond industry and corporate use cases, LEMs will eventually be able to map out accurate probabilities for future life outcomes and update those projections minute by minute. Han Solo once said, “Never tell me the odds,” but the average person can benefit from knowing and preparing for possible futures and get immediate feedback from the changes they make in their life. Think of how much faster you might be able to reach your goals with accurate event predictions by your side.