Retail’s AI Reality: From Experimentation to Execution
Retail and consumer goods are grappling with the opportunities and impact of mass consumer adoption of AI in 2026, and every industry conference is offering packed agendas of AI-related topics. Shoptalk Spring was no exception. There’s been a noted change in tone among brand executives, including more humility, recognition of uncertainty, and realistic reflection on how they are trying to operationalize AI in their organizations. It’s not just “AI is big.” Rather, it’s clear that retail has entered the implementation era of AI, and the brands that will win will be the ones who combine AI with human experience, clear business outcomes, and operational excellence.
We’ve evolved from our era of AI curiosity in 2024, through the AI hype of 2025, and we’ve arrived at the reality of AI in our businesses and the complexities it brings. Now we have to actually figure this out. The following themes reflect how leading retailers are navigating this shift and what it takes to turn AI from potential into performance.
From AI Exploration to Execution
Retailers are no longer asking “should we use AI?” They’re now mandating its use, prescribing the tools, and looking for help in selecting the most appropriate ones to make it work.
While AI may be new to some retailers, the approach to solving business problems hasn’t changed: start with the use case, quantify the ROI, and prioritize. This is comfortable territory for retailers. Few industries know their data better; test-and-learn is muscle memory, every KPI is measured weekly, and teams already know where the biggest revenue and cost-savings opportunities sit. That’s exactly where AI should be pointed first. In a Shoptalk Spring keynote, Salesforce shared that retailers running AI-powered marketing programs are generating 41% higher revenue per campaign than those operating manually.
Identifying where you can leverage AI is half the battle. Then comes the how, and this is where many of our clients stall, falling into the kind of deep analysis that delays the very progress that would justify the investment. We’ve seen teams assess tooling for months or pilot so many tools they never build any momentum. Given how fast AI tools like Claude are evolving, I’m partial to a Just Do Something approach: pick a problem with real business outcomes attached, choose a tool your team or partner is comfortable with, and ship a proof-of-concept. A working POC speaks for itself and gives leadership the evidence they need to invest more.
AI Is Reshaping Discovery, Not Replacing It
If there is a single phrase that defines retail in 2026, it's agentic commerce. Retailers are excited about AI agents that can browse, recommend, and even transact on behalf of consumers, but they are also being honest about how far the reality is from the vision. Bret Taylor of Sierra and OpenAI predicted at Shoptalk Spring that every retailer’s AI agent will eventually become as important as their website and mobile app. This hit home, as do his examples of what is working today, like the footwear brand whose agent lets customers photograph a damaged shoe and ships a replacement within 30 seconds, or Nordstrom rolling out an agent ahead of Black Friday.
AI is most powerful right now as a discovery engine, not a transaction engine. Consumers are using AI earlier in the journey to research, compare, and narrow options, and then validating with humans before they buy. As Gap’s CTO Sven Gerjets put it, “It’s not just keyword search anymore. It’s conversations, and so we need to be relevant to that.” Gap has launched conversational checkout and AI-powered fit recommendations while Sephora has rolled out its own ChatGPT app. The funnel is being rebuilt around AI-assisted discovery, but the human moment of decision still matters. For retailers, the immediate work is less about chasing fully autonomous agents and more about making your product data, content, and brand experience legible to the AI tools your customers are already using.
Operational Excellence Is Where AI Proves Value
The bar for AI investment has moved. In 2024 and 2025, “we’re piloting AI” was a credible answer in a board meeting. In 2026, it isn’t. The question every executive seems to be wrestling with is a simple one: what specific outcome is this AI application driving, and can I measure it? Innovation theater is out. Revenue, margin, and efficiency are in. The pressure is on to improve measurement strategies and demonstrate tangible ROI. PwC shares that only about one in eight CEOs report seeing measurable value from their AI investments today.
The fastest, most reliable returns are not coming from flashy customer-facing chatbots. Rather, they’re coming from operational AI in demand forecasting, inventory optimization, and logistics, where retailers reported double-digit reductions in overstock, stockouts, and fulfillment costs within months of deployment. We see in our client work that AI delivers the most when it’s pointed at problems where the data is rich, the KPIs are already defined, and the business case writes itself. If you’re trying to justify your next AI investment, start there. The flashier use cases can come once you’ve built the muscle and the credibility.
Physical Retail is Having a Strategic Resurgence
Physical retail is evolving. Stores are back, and not as legacy infrastructure, but as strategic assets. This shift is less about adding channels and more about clarity of purpose. New Balance offers a strong example. After a period of underperformance, the brand reset by defining exactly who it served and what it stood for, aligning decisions across product, pricing, partnerships, and store experience. They opened 80 new stores in 2025, a 150,000-square-foot Chicago flagship, and a real estate pipeline stretching five to seven years out. Physical retail became central to their strategy as a place to express the brand and build an emotional connection with customers. New Balance has grown 180% in five years on the back of that strategy.
The takeaway isn’t “physical beats digital.” It’s that the role of the store has changed. Stores are now where brands build emotion, demonstrate craft, and earn the trust that AI-mediated discovery can erode. Digital is for scale and convenience; physical is for connection and conviction. The retailers winning right now are designing their channels to do different jobs and then connecting the data behind them so the customer experience feels like one continuous relationship. Unified commerce isn’t a buzzword anymore, it’s the operating model.
Human + AI Collaboration is the Winning Model
The most quotable line of the Shoptalk Spring conference came from Reddit CEO Steve Huffman, whose keynote made a deceptively simple argument: people still want to hear from other people. “Coming out of the era of social media into the era of AI, there is a demand for more authentic and genuine places to engage and interact,” he said. Reddit’s own product strategy reflects this, their AI search product only responds in verbatim quotes from real users. Huffman noted that 40% of Reddit conversations are commercial because, in his words, “the question behind every question is, ‘What should I buy?’” Consumers are turning to AI for speed and to humans for trust, and the brands that understand that distinction are the ones who will stand out.
There is a growing strategic divide in how brands deploy AI. Some brands choose to show their AI and others choose to hide it. Brands speaking with the most conviction aren’t deploying public-facing chatbots, they're using AI invisibly behind the scenes while keeping the human moments of the experience front and center. As Sephora’s Global Chief Digital Officer Anca Marola puts it, “Intelligence equals trust.” AI scales decisions; humans provide the context, judgment, and emotion that turn a transaction into a relationship. The winning model isn’t AI or humans, it’s AI working underneath an unmistakably human brand.
Personalization at Scale is Now Table Stakes
Real-time, AI-driven personalization has shifted from being a differentiator to becoming the price of entry. Static segmentation is effectively dead. The brands at the front of the pack are personalizing creative, product recommendations, and full journeys in real time, using behavioral signals, purchase history, and contextual data that flow continuously across channels.
While AI finally makes the prolific goal of “personalization at scale” possible, retailers are hyper-aware that winning is in the execution. Personalization done badly pushes customers out of the experience; personalization done well makes the relationship deeper. If your personalization feels like surveillance, the brand has already lost. If it feels like the brand gets you and respects you, the brand earns loyalty in a market where the majority of consumers say they’ll leave a brand that doesn’t deliver it. Most retailers have the ambition, but are finding gaps in their plumbing. Real-time personalization requires unified data, fast decisioning, and creative systems that can keep up. That’s where the investment needs to go.
The New Tone is Humility
Compared to years past, executives are measurably less confident and noticeably more candid. We don’t have all the answers yet. It is evident that the industry is still figuring out what’s real, and that’s a healthy place for retail to be. The brands I’d bet on coming out of this era aren’t the ones with the loudest AI announcements. They’re the ones quietly doing the work by picking real problems, measuring real outcomes, blending AI with human experience, and accepting that this is a true transformation, not a quarterly initiative.
