Agentic Commerce Moves From Pilots to Production at NRF 2026

The National Retail Federation's annual show in New York made one thing clear: agentic commerce has crossed the chasm. NRF's first dedicated AI Stage drew packed sessions on agentic best practices, ROI, and personalized experiences, while Microsoft and Google both unveiled infrastructure plays that expand the addressable market for retailers.

Conversations with hundreds of retailers and dozens of sessions surfaced three defining trends, each showing a sector that has moved past the hesitation that marked last summer's sentiment and into active implementation and optimization.

Retailers Have Moved Past the "If" Question

The debate over whether agentic commerce is viable is over. The question now is how to deploy it at scale while preserving trust, brand identity, and control. Live audience polling during NRF's agentic commerce session found that nearly three-quarters of attendees were either actively implementing or planning agentic commerce initiatives.

That momentum is visible in the ecosystem. URBN, Etsy, Ashley Furniture, Coach, Kate Spade, Nectar, Revolve, Halara, and Abt Electronics have all onboarded to Stripe's Agentic Commerce Suite. More than 25 partners, including Salesforce, Squarespace, and PwC, have endorsed the Agentic Commerce Protocol (ACP).

Infrastructure-level announcements reinforced the shift. Stripe is helping power Microsoft's Copilot Checkout, which will let US-based Copilot users purchase from Etsy merchants, Urban Outfitters, and Anthropologie without leaving the chat interface. It's Stripe's second major AI agent integration, following its support for Instant Checkout in ChatGPT. Google countered with its Universal Commerce Protocol (UCP), adding another live standard to the market. For retailers already on the Agentic Commerce Suite, UCP support requires no extra integration work, since the suite automatically supports all agentic protocols through a single setup.

Agent-Ready Catalogs: Start Narrow, Then Expand

The most common question from retailers was consistent: what does good product data look like for AI agents? OpenAI's Yelena Reznikova, who manages B2B partnerships, stressed the importance of structured product feeds with clean, current item descriptions, pricing, and availability as the key to capturing intent and surfacing the right products at the right moment.

For retailers with enormous catalogs, that's a daunting prospect. URBN manages thousands of SKUs across its portfolio of consumer brands. The company's approach, as CIO Rob Frieman described it in an NRF session, was to avoid a big-bang overhaul. Instead, URBN zeroed in on high-impact categories like dresses and denim, standardizing language, attributes, and taxonomy in those areas first.

"You don't want to tackle everything out of the gate," Frieman said. "We have a very broad product catalog across all of our brands. We wanted to focus on some of our most popular products and use cases that were going to provide high value early."

First-Party AI Experiences Complement Third-Party Agents

Loyalty was a recurring concern among retailers: how do you maintain customer relationships when shopping happens through third-party agents? Many are answering by building their own native AI shopping experiences alongside their presence on external agent platforms.

Home Depot's Magic Apron, an AI companion available only on its own website, is one example. It handles specialized customer support and, because it has access to customer and purchase data that third-party agents lack, delivers a more personalized experience that builds on existing brand trust.

Ralph Lauren's David Lauren, chief branding and innovation officer, described a similar approach during his keynote. Ask Ralph, launched in September, creates shoppable outfit combinations from the men's and women's Polo collections, personalized to each user's prompts.

"Nobody needs a flannel shirt," Lauren said. "Showing a customer how to put it together is what makes us unique."

The takeaway from these examples is that retailers aren't choosing between third-party agentic discovery platforms and first-party AI experiences. They're building both, positioning themselves to meet customers wherever they prefer to shop while keeping deeper engagement on their own properties.