Prototypes, product feeds, and physical stores: My top takeaways from NRF APAC

As retail moves past AI prototypes and basic chatbots, it’s clear the industry is maturing. The strategic focus has shifted to autonomous agents capable of executing complex transactions, orchestrating multi-system workflows, and resolving back-end operational tasks. 

• The budget shift: As 47% of retail executives are now allocating over half of their entire AI budgets specifically to autonomous agents, their commitment to innovation is clear.

• The production hurdle: This sprint towards execution is urgent because, currently, only 25% of organizations have successfully migrated at least 40% of their AI experiments out of the pilot phase and into live production environments.

• The infrastructure playbook: Leading retailers recognize that long-term survival requires treating AI as a core architectural build-out rather than a temporary software plugin on top of old legacy setups. To stay agile, businesses must fundamentally rebuild their core frameworks to support an interconnected, agent-ready ecosystem.

For a deeper look at navigating this transition, check out our article on balancing AI discovery and conversion.

While over 90% of retail enterprises have actively launched data or AI initiatives, the vast majority remain completely stalled. Why? They lack the underlying operational foundation required to process, verify, or effectively use their data. The true bottleneck to scaling is not the AI software itself, but the fragmented, unverified master data sitting directly beneath it.

• The static data trap: Large Language Models (LLMs) inherently work with historical, scraped internet snapshots. Retailers must move past static data models and establish a live, real-time transactional system of record that accurately reflects actual in-stock inventory, exact regional pricing, and immediate fulfillment capabilities.

• The system disconnect: Advanced digital tools cannot succeed in isolation. Currently, roughly 50% of retailers operate with their front-end environments completely disconnected from their back-end infrastructure.

• The operational impact: Unifying these fragmented data channels across store locations, online apps, and phone orders optimizes corporate governance and gives a single source of truth. Also, it empowers your frontline store staff. By consolidating these systems, employees gain insight into a customer’s visit patterns and history data to deliver a highly tailored experience on the spot.

To see how you can unify your back-of-house systems, explore our full breakdown of why agentic commerce has an infrastructure problem.

The traditional consumer journey of moving sequentially from search to browse to buy is rapidly compressing into a crisp, three-step automated path: ask, decide, buy. Early ecosystem benchmarks highlight the sheer speed of this transition: we have already observed an explosive 8x growth in AI-driven traffic to structured digital stores alongside a massive 13x growth in orders originating directly from AI-powered search.

• The metadata shift: When discovery, product comparison, and checkout collapse into a single automated interaction managed by an external AI agent, a merchant's traditional brand storefront effectively turns into metadata for a machine to parse.

• Optimizing to get bought: The future belongs to brands optimizing for automated transactions over basic human search. This requires trading SEO for ACO by prioritizing machine-readable data and protocol-ready payment systems.

• The product feed gap: An immense operational gap exists today, as many retailers don’t have SKU structures ready to communicate natively with AI shopping agents. To win, brands must sanitize product metadata, maximize SKU accuracy, and ensure real-time inventory visibility so autonomous systems can

  surface and buy their items. 

• The human trust factor: Only 30% of consumers currently feel comfortable letting an AI agent complete a purchase on their behalf. However, consumer trust data explicitly shows that shoppers are three times more trusting of retailer-owned AI tools (25%) to manage their experience over generic, third-party AI platforms (7%).

This massive operational readiness gap is precisely why we launched Adyen Agentic. Built with key enterprise personas in mind, our framework resolves these exact front-to-back challenges through three core modules:

1. Agentic feed: Gives major AI platforms consistent, accurate product data with real-time pricing and stock information to erase errors that damage customer trust.

2. Agentic cart: Converts high-intent, AI-powered discovery into checkout-ready carts, with live pricing, tax, and shipping options automatically synced at the e

   xact point of machine selection.

3. Agentic payments: Accepts payments across all major agentic protocols and channels while staying compliant, secure, and fully in control.

Developed alongside partners like Google, OpenAI, and Meta, this infrastructure is already enabling tech-savvy enterprise retailers to lock in vital first-mover positioning in the AI agent economy.

To get your backend systems ready for this wave, review our agentic commerce pilot guide and our roadmap for preparing enterprise systems for autonomous commerce.

Only 30% of consumers currently feel comfortable letting an AI agent complete a purchase on their behalf.

Ben Wong

General Manager - SEA & HK

Despite the heavy, industry-wide focus on ecommerce, the physical store is still very much in the picture. Projections show that 75% of retail purchases will still happen offline by 2030. Shoppers still want to see, touch, and "hunt" for products in person, making physical stores a brand's experiential epicenter.

The limitations of AI analytics: While AI is highly effective for automating and optimizing the first 80% to 90% of the transactional or marketing discovery journey, the final 10% to 20% requires a dedicated human touch. Consumers buy products based on emotion, trust, and real relationship-building, so technology’s role should enhance human connection rather than simply add disconnected digital touchpoints.

Empowering employees: Instead of replacing human staff, real-time back-end data can elevate the store associate. Unifying payment and interaction data allows your on-the-floor employees to have more knowledge and control, like instantly recognizing whether a shopper is a regular or a first-time visitor, thus giving them a tailored experience.

A key message from NRF APAC was to avoid getting distracted by every shiny new piece of technology. Instead, focus long-term on the core operational challenges facing your business. The retail brands that thrive tomorrow will be those that treat technology not as an isolated experiment, but as a fundamental framework to empower their employees and serve customers.

Following our recent launch of Adyen Agentic, we are actively partnering with regional leading retailers to turn these strategic insights into actionable infrastructure. We help you establish the technical connections, protocol readiness, and security baselines required to take the friction out of automated commerce. 

Connect with our experts today to build an agent-ready payment framework.

Note: The insights and data points highlighted throughout this article are compiled directly from the keynote presentations, panel discussions, and industry case studies shared at NRF APAC.