If “data” was the dominant theme of the NRF APAC conference in 2025, this year it was “agentic.” Those terms are by no means mutually exclusive: A strong data foundation is crucial to unlocking real value from agentic AI.

At this year’s show, we witnessed an enormous amount of excitement around agentic AI and what it can do for customers. However, the deeper shift—the one that will separate winners from the rest—lies within the organization.

It takes more than technology to move from experimenting with agents to scaling them. Scaling requires redesigning operating models, workforce, and culture. Retailers must also address commercial and demand-generation implications, including agent discoverability, agent-ready production data, retailer-LLM partnerships, hybrid marketplace dynamics, and entirely new APIs and KPIs. 

Agentic commerce may be the most visible change in retail. The agentic operating model is the bigger one.

Here are our takeaways from NRF APAC 2026:

1. Everyone is talking about what agents can do. Almost no one is talking about how to scale them.

Agentic AI dominated the agenda at NRF APAC 2026, and we saw an impressive array of experiments and use cases. But there was a notable absence of any serious conversation about the organizational transformation required to move from a proof of concept to enterprise-wide value.

To close that gap, retailers must treat agentic AI as an operating model question, not a technology issue. That means rethinking how work is structured, how decisions are made, and, critically, how humans and agents work together.

The talent implications are significant. Retailers need new, AI-native talent and data scientists to partner with merchants and category managers, as well as new capabilities to design, manage, and optimize fleets of agents. Unlike the IT services boom, these skills must be built and embedded within the organization, not outsourced or bought off the shelf.

2. Readiness varies enormously, especially in talent.

The operating model is the central challenge; however, its intensity varies across the diverse market. We heard three distinct readiness stories playing out across the region:

• China: China has moved well beyond experimentation. Super-agent platforms are being used for everyday commerce activities—ordering products, navigating purchases, and handling customer interactions without a dedicated app or a human in the loop. Consumer trust in AI-powered tools is higher in China than in other markets, its technical capabilities are deep, and competitive pressures to scale are intensifying.
• Indonesia and the Philippines: In these markets, talent remains a key constraint. Retailers are struggling to find the agent architects, data engineers, and AI engineers needed to move at pace. Closing this gap requires sustained investment in technical education and capability-building across the region.
• India: Two contrasting scenarios are occurring in India. Some retailers are going “all-in” on AI priorities, focusing on scaling digital. Others are building out retail ecosystems more slowly, taking time to understand local consumers, build supply chains, and create a distinct value proposition. For the latter, AI is not the urgent issue—the fundamentals are. And they’re right. No amount of agentic sophistication will compensate for a weak value proposition, an uncompetitive assortment, inaccurate data, or disappointing delivery times.

3. Decision paralysis is just as dangerous as moving too quickly. Seek structured action, not a big-bang bet.

There’s a new source of anxiety gripping retail leadership: the fear of committing to a technology today, only to find something better available next quarter. AI capabilities are advancing at an extraordinary pace, and investment costs are still significant. Therefore, some retailers are choosing to quietly wait, deferring decisions and hoping the landscape becomes clearer.

However, waiting is not a neutral act. To scale agentic AI in two or three years, retailers must be building data foundations, operating muscle, and organizational capabilities now, even imperfectly.

Retailers cannot wait for perfect conditions, nor should they rush toward a big-bang transformation. This moment calls for disciplined action: Identify a specific team or business unit, run structured test-and-learns, and build toward scale through iteration.

4. Brand stories are becoming more important, not less.

Culture was an unexpected thread amongst all the tech-driven conversations. Leaders argued that the fundamental sources of retail differentiation remain unchanged: branding and consistent customer experiences.

For retailers, agentic AI is not just a channel or a cost lever. Agents are a manifestation of the brand and an extension of the culture. Before deploying any customer-facing agents, retailers should consider how they convey who you are.

What retailers should do now

The throughline from NRF APAC 2026 is clear: The technology is ready, but organizations are not. Retailers who pull ahead will not necessarily have the most sophisticated AI; instead, they will do the harder, less visible work of redesigning how their organizations function.

That requires starting small, taking controlled bets, attracting new talent, and building capabilities. Ultimately, they will ensure that technology is the latest chapter in their core brand story, rather than the story itself.

Agentic commerce is here. The agentic operating model must come next.