Q1. Could you start by giving us a brief overview of your professional background, particularly focusing on your expertise in the industry?

I’ve spent almost 15 years working in payments, beginning my career in retail banking and merchant acquiring at a leading local bank. That experience gave me hands-on exposure to how transactions work, how to manage risks, and how early tokenization frameworks were put in place. After that, I moved into the fintech world, where I headed up risk, legal, and compliance for a digital wallet startup in an emerging market. In that role, I focused on developing automated fraud detection systems to tackle complex threats, all while building out payment infrastructure to meet tough central bank requirements.

Later on, I took on senior leadership roles at a global payment network and its enterprise payment platform. My main goal was to drive digital transformation across different regions, manage large portfolios, and roll out advanced value-added services—like next-gen EMV tokenization, Click to Pay, and transit mobility solutions. I spent a lot of time working with top financial institutions, regional marketplaces, and major payment facilitators to help merchants increase acceptance, boost authorization rates, and improve overall payment performance.

These days, I work independently as an advisor to global consulting firms and expert networks. My main interest is where payments meet next-generation technology—especially how things like Generative AI and Agentic Commerce are set to transform the industry. I don’t see AI as just another efficiency tool; I believe autonomous AI agents will soon play a direct role in holding, initiating, and authorizing payments. That shift means we’ll need to rethink how we approach authentication, real-time risk management, and even the way payment systems are built.

Q2. In markets heavily dominated by zero- or low-interchange national real-time payment networks, what commercial or value-added approaches can international card networks deploy to prevent their lucrative card-not-present (CNP) e-commerce market share from being systematically eroded?

In markets where national Real-Time Payment (RTP) networks have little or no interchange fees, international networks can’t compete just by lowering prices. Instead, the smartest commercial strategy is to focus on bringing systems together and building out value-added services (VAS), based on three key pillars:

First, rather than going head-to-head with domestic Account-to-Account (A2A) payment systems, international networks should aim to work more closely with the local ecosystem. This involves teaming up with local banks, payment facilitators, and digital wallets to make international card credentials—especially credit lines—a core funding option right within the domestic RTP networks. By turning local QR code or real-time payment apps into distribution channels for credit, global networks stay relevant and can still capture high-margin credit transactions, even on what would otherwise be purely local payment rails.

Second, while national RTP networks are great at handling payments quickly and cheaply, they often don’t have the advanced risk management and dispute resolution systems that global card networks do. International networks can use this to their advantage by offering their enterprise-level value-added services to local players. By giving local banks and payment providers access to unified payment platforms and real-time fraud detection powered by machine learning, global networks can unlock new ways to make money from A2A data. Offering things like fraud prevention, tokenization management, and automated dispute resolution as subscription services means international networks can create steady revenue streams—even if the actual payment doesn’t go over a traditional card network.

Third, with digital platforms evolving so quickly—and new trends like Generative AI and Agentic Commerce on the rise—there’s a huge chance to rethink how digital payments work. Global networks need to step up their tokenization and digital authentication systems to support autonomous AI agents that use biometrics and digital credentials. By building these modern payment technologies directly into local banks’ online platforms through embedded finance, networks can offer instant, secure 'push provisioning.' That way, users can quickly and safely set up and use tokenized digital credentials in the new agentic commerce world—often before local A2A systems even catch up.

Q3. How has the bundling of business SaaS into smart terminal hardware altered the traditional pricing power and merchant retention margins of pure-play merchant acquirers?

Bringing industry-specific SaaS into Android-based smart terminals has completely upended the old-school model for merchant acquirers. Instead of just being a simple transaction tool, the point-of-sale has become a full-fledged 'operating system for business.' This change has reshaped payments in two major ways.

The first and most noticeable impact is that pricing power has made a comeback. In the past, merchant acquirers saw card processing as a commodity, which led to brutal price wars and shrinking margins. But now, when payment processing is built into business software—like inventory tracking, staff scheduling, or accounting—the conversation moves from 'What does this cost?' to 'How much value does this add?' Merchants are happy to pay higher, bundled processing fees because the software cuts down on admin work and helps their business run more efficiently.

But it’s not just about pricing power—this blend of hardware and software has completely changed how acquirers keep merchants on board. Traditionally, customer churn was all over the place, because merchants would switch to a new provider for even a tiny price cut. That’s no longer the case when the acquirer is also providing the software at the heart of a merchant’s daily business. Switching providers now means risking broken data, interrupting business, and retraining staff—all costly and disruptive. This deep level of reliance makes merchants far less likely to leave, leading to predictable, recurring SaaS revenue and steady transaction volumes.

Q4. When an autonomous AI agent holds and initiates a transaction with a biometric or digital credential, how must current card-not-present (CNP) risk engines and authentication rules adapt?

Agentic commerce is turning traditional CNP risk engines on their head—these systems were built on the idea that a real person is always present at checkout. But when an autonomous AI agent is the one making the transaction, risk engines and authentication systems need to move away from simply verifying the cardholder. Instead, they need a two-layered approach: one that checks the AI agent’s identity and another that verifies the intent behind the transaction using cryptography.

This shift starts with network tokenization evolving into more tightly controlled, agent-specific authorization frameworks. Instead of just using device-based tokens, global networks are now testing 'Agentic Tokens' that are directly tied to what the consumer wants to do. These new tokens, managed by emerging standards like the Model Context Protocol (MCP) or secure Web Bot Auth, go beyond hiding payment details—they build in programmable rules right into the transaction. That means the AI agent can be restricted to certain spending limits, merchant types, or timeframes. With these cryptographically enforced rules, risk engines can clearly tell the difference between legitimate automated agents and malicious bots or attackers.

But figuring out who the AI agent is only solves half the problem—authentication rules also need to confirm the human’s intent behind the machine’s actions. Legacy 3D Secure steps just add friction and don’t work when an AI is quietly handling transactions in the background. That’s why networks are teaming up with digital identity groups to test out 'Verifiable Intent' frameworks. Here, a person uses something like a biometric credential or a FIDO Passkey upfront to digitally sign a specific instruction for their AI. When the AI carries out a transaction later, it sends that secure, tamper-proof intent record along with the payment request. This lets issuer risk engines instantly review behavioral and contextual cues, skip any extra human verification, and approve the transaction smoothly based on the user’s pre-approved intent.

Q5. When deploying Account Funding Transactions (AFT) for instantaneous payouts, what are the primary risk-management, liquidity reserve, and AML compliance bottlenecks that local issuing banks face when handling these high-velocity, real-time push flows?

To really understand the challenges of Account Funding Transactions (AFT) and real-time push payments, it’s important to look at domestic and cross-border cases separately. In local, closed-loop setups—like B2C wage payments, digital wallet top-ups, or insurance payouts—the risk to local banks is pretty low. That’s because these payments come from well-regulated companies with clear funding sources and strong KYB/KYC checks, so they rarely put pressure on AML compliance or cause liquidity problems.

The actual bottlenecks emerge when these high-velocity push flows transition to cross-border networks for peer-to-peer (P2P) remittances, international freelancer payouts, or cross-border trade settlements. In this globalized context, local issuing banks face severe data fragmentation and operational strain in compliance. Because AFT flows move in seconds, real-time AML and sanctions screening engines are pushed to their architectural limits. The lack of standardized, structured data across jurisdictions leads to a higher rate of false positives. This forces local issuers to dedicate immense operational resources to manual review or to face high transaction-decline rates to avoid regulatory penalties for illegal cross-border layering schemes.

Compounding this compliance burden is a structural mismatch in international liquidity and settlement cycles. Even though an AFT push flow credit is reflected instantly in the end-user’s account, the net settlement between local banks and international card networks still operates on legacy, multi-hour, or next-day batch schedules. For local issuing banks handling heavy outbound freelance or P2P corridors across different time zones, this settlement lag creates immediate liquidity squeezes. To prevent localized cash shortfalls, banks are forced to maintain oversized, inefficient capital reserves with global correspondent banks. This capital inefficiency is further exacerbated by automated geofencing rules deployed by major networks and international payment facilitators, which often enforce hard declines on foreign-issued cards to comply with strict local capital controls, adding a layer of unpredictable payment friction.

Q6. When modern, scheme-led digital solutions are introduced, what are the primary structural and architectural roadblocks encountered within the legacy core banking systems of regional financial institutions?

The fundamental roadblock when introducing modern, scheme-led digital solutions into regional financial institutions is the irreconcilable friction between real-time, event-driven digital demands and legacy, batch-processing core banking architectures. Traditional mainframes were built decades ago to process transactions in static silos and overnight batches, whereas modern payment rails expect instantaneous elasticity and open data liquidity.

This structural disconnect becomes immediately apparent during the deployment of digital onboarding and card issuance frameworks. Modern digital issuance – enabling a user to undergo eKYC and instantly provision a virtual card – requires the core banking system to process unstructured data, validate digital identities, and sync credit lines in seconds. Instead, legacy cores struggle to update customer information files (CIF) in real time. When banks layer on network tokenization and push provisioning capabilities – such as forcing an instant 'Add to Apple/Google Wallet' stream from the mobile app – the legacy core often fails to digest the complex, modified metadata fields within the tokenized authorization messages. This inability to orchestrate real-time API handshakes with external token service providers leads to significant processing latency and directly dampens digital wallet activation rates.
Looking ahead, this architectural debt escalates from an operational bottleneck to a total system failure when confronting the paradigm of Agentic Banking. Autonomous AI agents operate 24/7 and generate high-frequency micro-transactions to execute programmatic machine-to-machine tasks. Legacy core systems, built with severely capped transaction-per-second (TPS) thresholds and rigid per-transaction cost structures, would instantly become paralyzed under the sheer volume of agentic micro-payments. Furthermore, traditional core security frameworks are entirely human-centric, designed to authenticate distinct user sessions via interactive OTP challenges. They lack the architectural capability to securely manage non-human, continuous machine sessions. Until regional institutions migrate toward modular, cloud-native microservices, their legacy infrastructure will remain a massive bottleneck, preventing them from accessing the next wave of AI-driven payment innovation.

Q7. If you were an investor looking at companies within the space, what critical question would you pose to their senior management?

If I were evaluating a company within this rapidly evolving payment and digital banking space, the single most critical question I would pose to their senior management is: 'How does your architectural model decouple transaction volume growth from capital inefficiencies and operational workarounds?'
Throughout the industry, we see many fintechs and regional financial institutions celebrating rapid digital adoption on the surface, while heavy operational dependencies lie hidden beneath. They are either relying on oversized, inefficient liquidity buffers to bridge legacy real-time settlement gaps, patching up obsolete core systems with fragile middleware, or applying human-centric risk rules to machine-driven transaction flows. 

As an investor, I want to know if a company is truly building a scalable, cloud-native infrastructure or just running a heavily subsidized operational illusion. Long-term, the ultimate winners in this space will not just be those who capture market share early, but those who can seamlessly integrate robust regulatory compliance, dynamic real-time risk engines, and capital-efficient settlements without collapsing under the weight of their own architectural debt.