GS interviews Ingenico's Erik Vlugt on what's next for AI in payments

As artificial intelligence moves deeper into the payments ecosystem, its role is shifting from back-end optimization to real-time decision-making at the point of transaction. That evolution is raising new questions around infrastructure, trust and the future of commerce. In this Q&A, Erik Vlugt, chief product officer at Ingenico, discusses where AI is delivering tangible impact today, what remains overhyped, and how the industry can balance innovation with the security and reliability that payments demand.

Green Sheet: AI is moving from back-office optimization into real-time transaction environments. Where do you see AI having the most structural impact in payments over the next three to five years, and where is it overhyped?

Erik Vlugt: AI will have its biggest impact in payments through real‑time decisioning. The most practical gains will come from better authorization performance, smarter fraud detection, and improved routing and device monitoring. These are areas where AI can make payments faster, more accurate, and more reliable.

AI will also influence how purchasing starts. As consumers use AI assistants to find products and make choices, payment systems must support transactions that begin through an agent, not directly with the shopper. This will test how well merchants stay visible and consistent across online and in‑store channels.

What’s overhyped is the idea that AI will replace the payments stack. Payments will remain regulated and trust‑driven. AI will enhance the infrastructure, not rebuild it.

GS: As payment devices and platforms become “AI-ready,” what does that mean at an infrastructure level? How should acquirers and large merchants think about AI integration beyond simple fraud scoring?

EV: Being “AI‑ready” means payment devices now act as connected, software‑based platforms. These devices link to merchant systems and the cloud, which allows AI to support better authorization decisions, smarter routing, and proactive device management.

For acquirers and large merchants, the value goes well beyond fraud scoring. AI can improve transaction performance and give clearer insight to device performance across large terminal estates. This shift also brings payments closer to broader commerce systems, making it easier to support new workflows and AI‑driven purchasing.

GS: We’re hearing more about agentic commerce and AI-initiated transactions. What new questions does that raise around authorization, liability, and proof of intent?

EV: Agentic commerce means AI assistants may start transactions on a shopper’s behalf. That raises core questions: how do we prove consumer intent, delegate authorization safely, assign liability, and connect agent‑driven online starts with in‑store fulfillment.

We will need stronger ways to confirm who is transacting and why. That includes clear consent and delegation, step‑up checks when risk is higher, and device‑level proof of presence or identity at the point of sale.

But this only works if data comes together. Transactions, device signals, risk events, and consent records must live in one place to be trusted and auditable. Without a unified platform, it becomes a non‑starter or a very high‑effort integration across many systems.

GS: From a risk and compliance standpoint, how does AI change the traditional model of certification, monitoring, and device lifecycle management?

EV: PCI and EMV remain the foundation. That does not change. AI adds continuous monitoring and adaptive risk controls, using real-time signals to spot anomalies earlier and manage large device fleets more proactively.

AI is also making development more efficient. Teams can ship faster with better code quality, quicker testing, and tighter release cycles. That helps time to market, but the security and compliance framework stays the same. Customers rely on us to stay secure and compliant. Ingenico owns the full stack from secure devices to cloud services, which gives us speed, control, and strong support. We use device telemetry and unified monitoring to improve uptime and compliance across markets, while keeping a stable certification baseline.

GS: AI promises personalization and operational efficiency - but payments is a high-trust environment. Where do you think the industry must draw firm guardrails?

EV: Payments runs on trust. AI should follow clear guardrails for security, privacy, and data governance.

Use payment data responsibly. Keep models and pipelines secure. Limit data access. Protect consumers. As AI drives more decisions, make those decisions explainable and auditable so banks, merchants, and regulators can see how risk is managed.

In times of change and opportunity, clients will rely on partners they trust. Ingenico has earned that trust over decades in card‑present payments. We own the key parts of the stack, from secure devices to controlled cloud services, which helps us safeguard data, enforce guardrails, and stay compliant. Our commitment is simple: innovate without compromising security, privacy, or transparency.

GS: Looking ahead to 2027 and beyond, do you expect AI to be a competitive differentiator in payments — or simply table stakes embedded into the infrastructure layer? What separates leaders from followers?

EV: AI will become a standard part of the payments infrastructure, but differentiation will depend on how effectively companies turn it into real improvements across the transaction flow. The leaders will be the ones who use AI to raise approval rates, reduce fraud, improve uptime, and keep experiences consistent across channels. This requires bringing together intelligence from devices, transactions, and merchant systems, and applying it both at the edge and in the cloud.

Execution speed will matter just as much as the technology. The companies that move quickly from proof of concept to full rollout will stand out, especially in areas like computer vision, preventative maintenance, and anomaly detection. The value comes from solving practical operational problems, not adding complexity for merchants.

Notice to readers: These are archived articles. Contact information, links and other details may be out of date. We regret any inconvenience.