NRF 2024: AI Stops Being Optional

Reading the coverage out of New York in January 2024, one theme was so dominant it crowded out nearly everything else. The National Retail Federation's Big Show had always functioned as a kind of State of the Union for commerce technology. Some years it confirms what you already suspected. This year, it confirmed something that felt like an actual turning point.

Artificial intelligence (specifically generative AI) was not a theme at NRF 2024. It was the entire conversation.

What Changed Between 2023 and 2024

The previous year's show had AI on the agenda, but tentatively. Retailers spoke about machine learning in the careful language of operational efficiency: better forecasting, smarter inventory, improved fraud detection. The technology was present but kept at arm's length.

Twelve months later, the frame had completely shifted. After eighteen months of ChatGPT reshaping public awareness of what AI can actually do, every major vendor and every retailer of scale arrived in New York with a generative AI story to tell. The question was no longer whether to adopt AI, but how fast and across which functions first.

Modern Retail's pre-show preview noted a shift in tone from previous years: rather than focusing on macro forces outside retailers' control (the supply chain crisis, inflation, the pandemic hangover), the 2024 programming was oriented toward practical use cases. Execution, not aspiration.

The Numbers Behind the Shift

The most striking data to come out of the show's opening days was Nvidia's first annual State of AI in Retail and CPG survey, released just before the event opened. Forbes reporter Joan Verdon, covering the opening day from the Javits Center, reported its headline findings: 76% of surveyed retailers were already using AI or actively preparing to pilot it.

That is not a marginal number. And it was accompanied by a note of genuine surprise from Nvidia's own analyst: "The adoption is actually farther along than we thought," said Cynthia Countouris, Nvidia's Director of Product Marketing for Retail, CPG and QSR.

Already using AI or preparing to pilot

76%

Planning to expand AI infrastructure investment

64%

Credit AI with increased annual revenue

69%

Report decreased operating costs from AI

72%

The same survey found 64% of retailers planned to expand their AI infrastructure investment over the following 18 months, with 34% of those planning to increase spending by more than 15%. The investment intent was not modest.

The top five active AI use cases among those surveyed were: personalised customer recommendations, store analytics and insights, adaptive advertising and promotions, stockout and inventory management, and conversational AI. Roughly the full stack, in other words: from back-of-house to customer-facing, from prediction to generation.

The Platform Plays

Post-show reporting suggested the headline announcements had come from the major platforms.

Salesforce used the show to announce Einstein Copilot for Commerce: a conversational assistant embedded directly into its Commerce Cloud product, designed to let shoppers describe what they're looking for in natural language and receive personalised results. The pitch was that the technology had moved from experiment to product.

Microsoft expanded its Copilot integrations for retail, leaning into its partnership with OpenAI to position the stack as AI infrastructure for mid-market and enterprise retailers. The framing was integration depth: AI that augments existing systems rather than requiring retailers to rebuild from scratch.

Google's presence, according to reporting from the show, was focused on search. Specifically, the argument that AI-powered discovery is about to reshape how customers find products. With generative search experiences arriving in Google's core product, the implication for retailers was significant: SEO strategies built over the preceding decade may need rethinking.

None of that will surprise anyone who was watching the technology landscape in late 2023. But hearing it at NRF, from the platforms with the largest commercial relationships in retail, marked something. It shifted the conversation from "will this matter" to "when and how."

On the Shop Floor

Away from the keynote stages, coverage from the exhibition hall documented several recurring themes.

AI-native customer service was perhaps the most active space. Multiple vendors demonstrated systems capable of handling not just simple FAQs but complex, multi-turn interactions (returns, order modifications, product queries) without human intervention. The accuracy claims ranged from credible to optimistic. The direction of travel was clear.

Computer vision for retail operations had a noticeably stronger showing than in previous years. Applications included automated inventory counting (cameras scanning shelves and flagging gaps in real time), checkout friction reduction, and loss prevention. Forbes coverage noted that exhibitors were demonstrating how computer vision and AI analytics "can gather information about how consumers shop, and can be used for better store operations, merchandising, layout, and stock management." The physical store, long AI's poor relation to e-commerce, was starting to catch up.

Personalisation at scale, historically one of retail AI's most-discussed and least-delivered promises, seemed genuinely closer to reality. The generative AI step change had opened up dynamic content generation that previous ML-based approaches couldn't match in cost or flexibility.

The Honest Counterweight

To NRF's credit, the 2024 show made room for scepticism alongside the enthusiasm.

Coverage from practitioner panels described a consistent pattern among retailers who had spent 2023 running GenAI pilots. The technology had crossed a threshold. But the implementation work was harder than the demos suggested. Data quality, integration complexity, change management, the difficulty of measuring short-term ROI: these were the unglamorous realities that vendor keynotes tended to skip. Several retail executives were candid that moving from pilot to production was proving slower than expected.

That honesty is useful. The Nvidia survey figures are self-reported, and "preparing to pilot" is not the same as "in production." The gap between intention and deployment is where most enterprise technology programmes lose momentum, and retail has a particular talent for accumulating impressive-sounding pilots that quietly expire.

What It Means Looking Forward

Reading the post-show analysis, NRF 2024 felt like the moment AI moved from strategic initiative to operational necessity for commerce. Not because every retailer was deploying sophisticated AI systems in January 2024. But because the conversation had irreversibly shifted. The question had changed.

For UK retailers watching from a distance, the show was a useful calibration. The US enterprise market was moving faster, with more capital and more vendor options. But the same underlying dynamics applied on both sides of the Atlantic: ChatGPT mainstreaming consumer expectations, foundation model costs dropping, the data quality imperative becoming inescapable. The implementation challenges being reported from New York were the same ones UK retailers were encountering in their own early pilots.

The retailers who will benefit most are those who use the coming period to build the foundations: clean, unified data; clear use-case prioritisation; and the organisational capability to iterate on AI applications as the technology and their understanding of it matures.

The retailers who will struggle are those treating AI as a feature to bolt on rather than a capability to develop. The gap between those two postures was already visible at NRF 2024.

---

NRF's Big Show runs annually in New York each January, at the Javits Center. The 2024 edition, held 14–16 January, was expected to draw close to 40,000 attendees from across the global retail and technology industry.