This article is provided by BRC Associate Member, Ecommpay.

__________________________​

The way we shop is evolving, and AI is at the heart of the change. Agentic commerce – where intelligent agents act on behalf of consumers to select, purchase, and pay for products – is being described as the next seismic shift in e‑commerce, rivalled only by mobile. But amid the excitement, one fact is often overlooked: the payments infrastructure needed to make this possible largely already exists.

Advances in tokenisation, authentication standards, payment orchestration, and regulations such as PSD2 and PSD3 have created a framework for secure, delegated payments. From a fraud perspective, it doesn’t matter whether a human presses “buy” or an AI does it – the same core protections apply.

Different models, different futures

The real question isn’t whether agentic commerce can happen – it can – but which approach will take the lead.

Some systems are tied to specific platforms, creating closed ecosystems where only selected merchants can participate. Shopify’s ShopPay illustrates this model. Other approaches are browser-based, designed to interact with a variety of payment methods across multiple merchants, similar to today’s search engines. Google or Perplexity could feasibly develop in this way.

The implications are profound. Will consumers notice – or care – if their AI agent only recommends certain merchants? And what does this mean for retailers outside those ecosystems whose payment methods may not be prioritised?

Trust is still the cornerstone

Trust has always underpinned payments. In an AI-driven world, the speed and scale of decisions are accelerated. Intelligent agents will assess merchants in milliseconds, factoring in data such as delivery reliability and payment performance. For retailers, this makes it crucial that infrastructure supports high approval rates, smooth authentication, and minimal friction at every stage.

Emerging standards like Know Your Agent (KYA) aim to verify that AI is acting on behalf of genuine, consenting users. Yet new threats are emerging – from synthetic identities to AI-enabled scams – demanding constant vigilance alongside familiar fraud defences.

Preparing for change without upheaval

Agentic commerce won’t arrive all at once. Its adoption will be gradual, with competing models testing both consumer trust and retailer readiness. Businesses don’t need to rebuild their systems from scratch, but they do need to ensure their payment infrastructure is resilient, flexible, and ready to integrate with AI-driven workflows.

This involves:

  • Designing payment systems that work seamlessly with AI agents
  • Maintaining high approval rates through smart routing and local acquiring
  • Making offers machine-readable so agents can quickly verify and act
  • Understanding how AI platform preferences can influence consumer choices

Looking to the horizon

Agentic commerce isn’t mere hype, but it isn’t a wholesale reinvention of payments either. The foundations – secure, consent-based payments – are already in place. The real challenge will be which model dominates: closed, platform-specific ecosystems or open, interoperable networks – and how retailers position themselves within that landscape.

Those who succeed will be the ones who keep their payments infrastructure agile, stay informed about evolving AI platforms, and adapt to changing consumer expectations. Agentic AI also has the potential to make e-commerce more inclusive - guiding consumers through checkout and payments, enhancing accessibility, and helping those who might otherwise struggle to complete transactions online.

Ultimately, the biggest shift may not be in technology itself, but in who controls access to the consumer.

Article provided by