Transaction Trends

The Big Picture

AI in payments is moving beyond incremental efficiency gains toward a transformation of the industry’s underlying global operating system. What began as a set of tools focused primarily on fraud detection and risk management is evolving into agentic systems capable of autonomously managing pricing, routing, and interchange outcomes across merchant portfolios, while also enabling consumer-facing agents that can browse and transact on a user’s behalf.

The shift marks a fundamental change in how payment decisions are made and executed across the ecosystem. As AI-driven systems take on more responsibility for real-time optimization and decisioning, their impact extends well beyond user experience to the economics, governance, and resilience of payments infrastructure.

As Matt Downs, President of Integrated & Platforms at Global Payments Inc. (Worldpay), has described this next phase of AI in payments: “Horizon two will usher in the next generation of platform products where AI isn’t just an add-on but the foundation. It will assist users and developers in real time and even participate in building the product itself in a highly adaptive and intelligent way.

Zooming in: Agentic Commerce

Agentic commerce refers to AI systems that can autonomously make and execute decisions within payment flows, rather than simply analyze data or follow static rules. These systems act within human-defined objectives, constraints, and policy guardrails, executing decisions at machine speed while remaining subject to oversight and control.

While consumer-facing agents that browse and transact on a user’s behalf are one expression of agentic commerce, the near-term impact is likely to be greatest in backend payments infrastructure, where autonomous decisioning can materially affect cost, performance, and resilience.

With the help of payment service providers, AI systems are beginning to participate directly in commerce flows, both at checkout and behind the scenes. The following scenarios illustrate how agentic commerce is emerging across the payments ecosystem.

• Agent-assisted checkout (consumer-facing): A shopper uses an AI agent to discover a concert in their hometown. The agent supports the checkout experience by surfacing ticket options and preparing payment details. The shopper confirms the purchase, and the transaction is completed using tokenized credentials through the merchant’s payment service provider.
• Autonomous purchase execution (consumer-facing): A shopper authorizes an AI agent to make a future purchase within defined parameters (e.g., securing two side-by-side concert seats under $300 once tickets become available). The agent monitors availability and executes the transaction at the right time without further user involvement.
• Dynamic routing optimization (infrastructure-level): For a multi-merchant platform, an agentic system evaluates real-time authorization performance, network costs, and merchant preferences to determine the optimal routing path for each transaction. Rather than relying on static routing rules, the system continuously executes routing decisions within human-defined constraints to balance approval rates and cost of acceptance across the portfolio.
• Autonomous interchange and pricing management (infrastructure-level): An agentic system monitors merchant interchange performance against predefined pricing tiers in the merchant’s contract and can automatically apply the appropriate tiered pricing within processor-defined guardrails, reducing manual intervention while improving margin consistency.

Taken together, these scenarios demonstrate that agentic commerce extends beyond consumer convenience. By automating complex, high-volume payment decisions, its most immediate impact may be changing how pricing and routing decisions are executed and how interchange performance is managed across the payments ecosystem.

The Payment Ecosystem’s Role

Recent Activity

Payments players are laying the infrastructure for AI-first payments, applying secure tokens that mask credentials, biometric authentication, and transaction controls to guard purchases.

Stripe and OpenAI launched the Agentic Commerce Protocol to enable instant checkout in ChatGPT, starting with Etsy and Shopify. This protocol also works across payment processors.

Visa and Mastercard have been establishing guidelines for agentic commerce, developer standards, and partnerships with companies such as Elavon, Coinbase, and Microsoft. On the infrastructure side, progress toward agentic autonomy is emerging more quietly. Rather than new protocols, it is being enabled by richer post-clearing data, portfolio-level analytics, and pricing frameworks that allow merchants to move between predefined tiers based on observed interchange performance. Together, these capabilities create the foundation for agentic systems to execute pricing and routing decisions at scale within established contractual and regulatory guardrails.

Action Areas

1. Define where to partner versus where to differentiate.

Payments companies should be explicit about which parts of the AI stack they intend to build and which they will access through partnerships. AI platforms and payment networks are rapidly defining standards for agent-initiated commerce, but value creation for processors, ISOs, and acquirers will increasingly come from how those capabilities are applied to routing, pricing, and portfolio management. Clear partnership strategies help avoid duplicative investment and ensure faster time to market.

2. Treat merchant and developer experience as an economic lever, not a design exercise.

As AI becomes part of the payments value proposition, onboarding and documentation quality directly affect adoption and realized revenue. If a payments player wins business based on AI-enabled simplicity or automation, the onboarding experience must reinforce that promise. Payments-focused analytics, consulting, and enablement firm TSG conducts benchmarking analysis that shows onboarding APIs average just 47% in overall quality, indicating a material execution gap. Quality is measured against market best practices for merchant enrollment, record updates, and risk data. Scaling AI-driven capabilities depends on improving documentation quality, expanding testing environments, and standardizing integration workflows.

3. Prepare to support software partners as AI adoption accelerates.

Software platforms increasingly view payments as core to their business. In a joint survey from TSG and Fiserv, over 90% of software companies identified payments as critical to their success, yet only a minority viewed their current provider as a strategic partner. As AI-driven use cases expand, payment players will need to support software partners with tools, data access, and operational flexibility that enable agentic features without increasing complexity or risk. Payment providers that can bridge this gap stand to deepen partner relationships and capture greater long-term value.

4. Modernize architecture to enable agentic execution, not just analytics.

Agentic systems require more than insights; they require systems capable of executing decisions reliably within defined guardrails. Disconnected platforms, brittle data pipelines, and manual intervention points limit the ability to automate at scale. Whether through system consolidation, AI-native platform development, or accelerated integrations, infrastructure modernization is an imperative for meaningful AI adoption.

Challenges

As payments become more automated, fraud risks continue to evolve. In November, Visa reported a surge in dark-web discussion on AI agents, as well as a 25% increase in fraudulent transactions initiated by bots over the prior six months, with the U.S. experiencing a 40% rise.

While bot-driven activity is not synonymous with AI-enabled payments, the trend underscores the need for strong authentication, monitoring, and controls as software-driven interactions become more common across commerce flows.

Separately, AI systems trained on biased datasets may unintentionally favor certain merchant types, cardholders, or transaction patterns, resulting in unfair outcomes that erode consumer trust and brand identity. Finally, as AI-enabled systems take on greater responsibility within payment flows, questions around accountability, transparency, and liability for errors become increasingly important. Addressing these issues will be critical to sustaining trust as AI-driven payments continue to scale.

The Bottom Line

The payments industry has always demonstrated a remarkable ability to manage threats, drive innovation, and power the global economy. Major advances are already underway, pushing commerce toward greater levels of seamlessness and interoperability. The next phase of that evolution will be defined less by new checkout experiences and more by how effectively the industry applies AI to automate complex, high-volume payment decisions.

Further, trade associations such as ETA have prioritized fostering a policy environment that protects consumers while allowing competition and innovation to thrive, as reinforced by Jodie Kelley, CEO, ETA in her January testimony in front of The House Financial Services Committee Subcommittee on Digital Assets, Financial Technology and Artificial Intelligence and with AI identified as a key topic in the Association’s 2026 Federal Legislative and Regulatory Policy Priorities.

About ETA

The Electronic Transactions Association (ETA) is the world’s leading advocacy and trade association for the payments industry. ETA members span the breadth of significant payments and fintech companies, from the largest incumbent players to the emerging disruptors in the U.S. and in more than a dozen countries around the world. ETA members make commerce possible by processing approximately $56.75 trillion annually in purchases and P2P payments worldwide and deploying payments innovation to merchants and consumers. ETAs membership spans the breadth of the payments industry to include independent sales organizations (ISOs), payments networks, financial institutions, transaction processors, mobile payments products and services, payments technologies, and software providers (ISV) and hardware suppliers. For more information, visit electran.org.

The Electronic Transaction Association brings leaders from across the payments ecosystem together for collaboration, advocacy, and education. ETA members are leading the charge, leveraging AI to drive efficiency, expand market reach, and deliver smarter payment experiences. As these innovations accelerate, ETA is championing policies for responsible AI adoption and fostering collaboration through its AI Committee and industry events such as TRANSACT, Strategic Leadership Forum, and the Payments Compliance Conference. Mark your calendar for ETA’s TRANSACT event in March to engage with AI-focused content and stay ahead of AI in payments.

About TSG

TSG (The Strawhecker Group) is a globally recognized analytics and consulting firm that supports the entire payments ecosystem, serving over 1,000 clients from Fortune 500 leaders to more than a dozen of the world’s most valuable brands. Trusted by industry leaders, TSG’s strategic services, market intelligence, and analytics merge to empower clients with actionable and accessible information.