Veteran retail technology expert, Jon Stine, shares his reflections on what NRF’s crowded expo floors told us about how AI is being deployed and how it is remodeling retailing.  

The more than 40,000 attendees at the mid-January National Retail Federation Big Show in New York might be excused if they came away believing that the word “retail” was always followed by the term “artificial intelligence.” 

AI – and claims of AI functionality – was prevalent on every level of the Expo Floor, a common topic in business-focused presentations, and the theme of multiple well-attended off-site gatherings. 

A sifting of the pitches, panel presentations, and PowerPoint leads to these summary perspectives on artificial intelligence in retail as we enter 2026: 

• The industry’s artificial intelligence conversation is now shifting from technology to value.  We saw – appropriately, in our minds -- a growing insistence on operational impact (and better yet, ROI based on a deep understanding of business process and operational metrics.)  Said one leading retail analyst: “there was much less interest in shiny objects – people were getting real.”     

• Artificial intelligence is not one technology, but several, each with unique capabilities and value. Leaders are implementing what might be termed an “AI cocktail,” a workflow-guided mix of machine and deep learning (for correlation-based prediction), computer vision (traffic and aisle analysis), natural language processing (for customer service and multi-language task management), spatial and physical AI (VR, AR, and robots), generative AI (for individual productivity, research and early-iteration creativity), and now, agentic artificial intelligence.  Looming over the horizon: causal AI.   

• The majority of the retail industry is in its early phase of generative and agentic AI maturity.  Keep in mind that AI in retail is not new; more than 40% of the world’s top 100 retail brands were using machine learning AI in 2017, and natural language AI use in retail call centers has been common for years.  However, the bell curve of next-generation AI adoption reflects a bulge of firms that have moved beyond tech pilots to process-centric experimentation – but have yet to realize desired (or promised) gains.  The bell curve also shows a few – equipped with talent, budget, and business strategy – racing ahead and harvesting value, and a long tail of stragglers.   
   

• Both definitions of agentic artificial intelligence were headline topics at this year’s NRF Big Show – and both will be at the center of strategic industry discussions through 2026.  One, otherwise known as agentic conversational commerce, is about consumer use of generative AI large language models (LLMs, such as ChatGPT) to search for, discover, and purchase retail products and services.  One report suggested that agentic conversational commerce influenced more than 20% of all online retail sales globally between 1 November and 31 December.  The second is about the use of artificial intelligence “agents” (understood as autonomous, task-specific systems that can acquire, analyze, and act upon data, and work in concert with other agents) to automate and speed operational workflows.   

Adoption of agentic conversational commerce is expected to increase rapidly among shoppers worldwide, with some studies suggesting that 50% or more of global online shoppers will regularly begin their 2026 decision journeys through a LLM.  Indeed, artificial intelligence may now (and properly, in our minds) be positioned as a primary brand interface.  

Adoption of agents within operational workflows will move more slowly. 

• The dependencies of AI still loom large. These are issues at the heart of successful production implementations, and are the no-man’s-land between pilots/POC’s and business value: 

• Data quality and accessibility.  This, the great barrier that prevents many brands from getting from here to there.  If ignored through the years, remediation can be painful – but it is necessary. 

• Change management. Technology is but one (and a minority) part of AI implementation, especially as the industry begins its agentic transformation of operational workflows.  How will processes, decisions, and next-step actions change?   

• The human element.  Inside the store, the distribution center, and headquarters, AI will re-shape employment.  The question for brand leaders is how; will it be attrition (AI used to cut headcount) or augmentation (AI used to make employees smarter-faster)?    

One thing is certain: artificial intelligence will again be on the agenda come January 2027. 

Jon Stine is Co-Founder of Honeycomb Retail AI