Introduction

Coresight Research is a research partner of Shoptalk Spring 2026, taking place March 24–26 at the Mandalay Bay resort in Las Vegas, Nevada. In this report, we present key insights from the first day of Shoptalk Spring 2026, which covered three themes:

• Agentic Commerce and AI-Driven Personalization—Retailers and platforms are using AI to move from reactive product discovery to proactive, conversational shopping experiences managed by intelligent agents. However, agentic commerce remains in early stages, limited by gaps in infrastructure such as universal carts, APIs and standardized product data, with the strongest near-term impact seen in search, merchandising and supply chain optimization.
• Retail Media Innovation and Omnichannel Data—Retail media is increasingly driven by first-party data, enabling more precise targeting and personalized engagement across diverse customer segments. At the same time, improved ad tech integration is connecting media exposure to actual transactions, allowing brands and retailers to measure performance and optimize campaigns at a granular level.
• Cultivating Emotional Connection and Loyalty—Effective merchandising balances data with merchant instinct, using customer and demand signals to identify patterns and opportunities while relying on human judgment to interpret what those signals could become. As AI-generated content grows, consumers are increasingly turning to peer-driven, community-based sources for the nuanced, experience-led validation that algorithms cannot provide.

Coresight Research Insights

1. Agentic Commerce and AI-Driven Personalization

Track Keynote: Using AI to Amp Up Personalization

AI is changing how personalization works, moving discovery from text-based search to visual, intent-led exploration. Matt Madrigal, Chief Technology Officer at Pinterest, said, “Taste is visual—you know it when you see it, even if you can’t describe it.” Pinterest positions itself at the earliest stage of the shopping journey, capturing intent before purchase decisions are explicitly formed. By reading visual signals, metadata and behavioral patterns, the platform constructs a “taste graph” to anticipate user preferences and guide them toward shoppable outcomes.

Half of the platform’s users are Gen Z, and 39% of them now use Pinterest before turning to Google. From a Coresight Research perspective, this points to a broader shift: personalization moving from reactive recommendations to proactive inspiration, driven by AI-native platforms.

Matt Madrigal, Chief Technology Officer, Pinterest
Source: Shoptalk Springl

 

Retailers are shifting from experimenting with AI to embedding it into core operations and customer experiences. Sven Gerjets, EVP & Chief Technology Officer, Gap Inc., noted that Gap is taking a structured approach through its “Office of AI,” focusing on organizing workflows and decision-making around the customer journey rather than simply automating tasks. The company is applying AI in high-impact areas like product discovery, checkout and sizing while also rethinking data and workflows at scale.

Gerjets stressed the need for speed given how rapidly AI is evolving, which drove Gap’s early partnership around Google’s UCP (Universal Commerce Protocol), designed to let merchants bring their own experiences into LLM-driven commerce journeys. For Gap, the value is not just visibility inside AI interfaces, but preserving merchant control over relevance, customer data and the end-to-end brand experience. Gerjets said, “We also don’t want to push our customers out of the experience. We want to maintain the relationship.”

Sven Gerjets, EVP & Chief Technology Officer, Gap Inc (Left); and Gabrielle Fonrouge, Retail Reporter, CNBC (Right)
Source: Shoptalk Spring

 

Track Keynote: Building a Next-Generation Shopping Experience

The launch of new shopping capabilities marks the early stages of agentic commerce, shifting from product-led search to intent-driven discovery. Mahak Sharma, Product Partnerships, Search & Commerce at OpenAI, described how consumer behavior is changing: users now come with broad needs and constraints, looking to solve problems rather than searching for specific product links. The platform is launching flexible merchant pathways that allow checkouts both on- and off-platform, adapting dynamically to the user’s mission.

The interface is changing, but retailer differentiation still relies on consumer confidence. Anca Marola, Global Chief Digital Officer at Sephora, said, “Intelligence equals trust.” Sephora is building out its customer intelligence to create a conversational flagship experience online, extending its role as a trusted beauty advisor into emerging platforms. This introduces a real tension between platforms positioning themselves as the primary discovery interface and retailers trying to preserve their brand authority and customer relationships.

Left to right: Anca Marola, Global Chief Digital Officer, Sephora; Mahak Sharma, Product Partnerships, Search & Commerce, OpenAI; and Hilary Milnes, Executive Editor, Vogue Business
Source: Shoptalk Spring

 

The New Retail Paradigm: Agentic Commerce & Supply Chain in the AI Era

Leading brands are pointing AI models at enterprise data to overhaul supply chains and operations. Ryan Ball at Anthropic noted that the state of AI changed six months ago, moving from question-and-answer assistants to agents capable of deep research and autonomous work.

David Hardiman-Evans at Ocado put it plainly: fully automated grocery fulfillment through robotics would not be possible without the leap AI represents. Nitin Murali at Gallo built supply chain and demand intelligence applications with zero coding experience and surfaced insights the company had never seen before. Murali noted that all of it was built using Claude.

Building and Deploying AI Agents

The best-performing AI agents will be specialized and coordinated, with humans managing humans, agents and human-agent teams. Jason Cottrell, CEO at Orium, said enterprises need to use both internally developed and external technology expertise, because they cannot easily replicate the product of entire companies developing a single platform. Robert Ibarguen, Head of Digital Intelligence at Newell Brands, noted that a single generative AI project can spin off benefits and insights for teams well outside its original scope. For example, an automated sentiment engine for ratings and reviews evolved to provide action-oriented insights for innovation and other teams. Data gaps are likely inevitable, but synthetic data can help bridge them.

Track Keynote: Sorting Agentic Hype from Reality

Agentic technologies are set to become the new digital front door for retailers, bringing telephone and online customer support under one roof. Bret Taylor, Co-Founder and CEO at Sierra, and Chairman of the Board at OpenAI, predicted that every retailer’s AI agent will eventually become as important as their website and mobile app. Taylor said that forcing customers to wait on hold is disrespectful, and that deploying faster, cheaper AI agents for functional inquiries frees up budget for human-touch experiences. Taylor also said AI will see more use for intimidating, considered purchases, and warned that if an AI platform provides a better discovery experience, consumers will bypass the retailer’s website entirely.Taylor pointed to deployments already under way. Nordstrom launched an AI agent ahead of Black Friday; Taylor also described a footwear retailer whose agent allows customers to photograph damaged shoes, adjudicates the claim instantly and ships a replacement within 30 seconds. These, he argued, are the moments where brand loyalty is actually built or lost.

Perspectives from Investors on Tomorrow’s Crucial Retail Technologies

Investors are highly selective in the current funding environment, with a small number of companies attracting capital while many find it harder than ever to raise. Matt Nichols, General Partner, Commerce Venture, flagged tools that require human action to demonstrate return on investment (ROI) as a red flag, favoring AI that enables autonomous systems of action. According to Janie Yu, Partner, LFX Venture Partners, red flag is solutions built as mere wrappers around LLMs with no proprietary technology, while domain expertise and genuine problem understanding are what draw her in.

On agentic commerce, both agreed the infrastructure is not yet there. The most critical missing piece, according to Nichols, is a universal API to accept agentic orders—and Yu added that any cart solution must be both multi-merchant and multi-item. Underpinning all of it is product data. A universal SKU identifier—able to recognize the same product across retailers regardless of how it’s described—was cited as essential infrastructure for agentic commerce to scale.

In terms of where AI is delivering real value today, investors see the strongest ROI in areas such as search, merchandising and supply chain optimization. AI-driven search, in particular, was highlighted as an area of meaningful improvement. By contrast, fully autonomous shopping experiences remain overhyped, with consumers still valuing discovery and decision-making over end-to-end automation.

Startup Pitch: Technologies Reinventing Customer Experience

Startups are building the tools this agentic future will need. In the Startup Pitch, Deborah Weinswig,CEO and Founder at Coresight Research, noted that companies can “use AI to pay for AI” by eliminating existing costs and frictions. The Startup Pitch saw 10 innovators pitch for the Judges’ Choice and Audience Choice winners, with the former receiving an optional $100,000 in funding from Commerce Ventures and the latter receiving $100,000 of research and services from Coresight Research—see our Startup Pitch preview report here.

Max Sinclair, Founder & CEO at Azoma, introduced an Agentic Merchant Protocol designed to help brands get recommended more frequently on AI agents, combating the shrinking digital shelf. Rajan Balasundaram, CEO at Bayezon AI, showed a platform that connects intent, context and product data to bypass the current consumer browsing model. Vita Mallela of Flock AI showed a custom-built tool for fashion brands that creates on-brand content at one-tenth the cost of a traditional photoshoot. Ben Seidl, Co-Founder and CEO at Autolane, presented a hardware-enabled software platform that positions autonomous vehicles as the delivery infrastructure of the future, using a smart curbside sign to coordinate across self-driving platforms and connect them to retailers, restaurants and shopping centers—projecting 10 million autonomous vehicles on the road by 2030. Anna Liang, Founder and CEO at SocialQ AI, addressed the social commerce challenge for small and midsize businesses, presenting a platform that creates on-brand content and publishes it automatically, delivering what Liang described as agency-level outcomes for brands that lack the budget for an agency.

Startup Pitch: Technologies Driving Operational Efficiencies

A second session of the Startup Pitch focused on back-of-house operations. Weinswig noted that, at Coresight Research, we are covering how AI can help retailers achieve greater productivity in retail—see our new report here—and that this theme was relevant for a number of the participating Startup Pitch companies. Weinswig noted that retailers should consider how they are sequencing for AI productivity gains and where they are investing.

Coframe, the only technology company with a direct partnership with OpenAI, presented an AI-driven conversion rate optimization platform that enables websites to test and optimize themselves continuously. Ethosphere introduced microphone technology for store environments that analyzes associate-guest interactions, generating personalized coaching for associates, management tools for store managers and brand-level insight into what drives sales on the floor. MerchKit addressed the product discovery challenge posed by agentic commerce, building a catalog platform that continuously audits product content against the specific attribute requirements of individual AI platforms and LLMs. RateRunners targets shipping cost leakage, checking every shipment against carrier invoices to catch billing errors and recover disputed charges. ShopSight presented a demand forecasting platform that co-creates new product concepts with customers in real time, allowing brands to test demand before committing to a production run.

The winners of the Startup Pitch competition were:

• Judges’ Choice: Coframe
• Audience Choice: Flock AI

2. Retail Media Innovation and Omnichannel Data

Retail Media Innovation: Retailer + Brand Case Studies

Commerce media runs on first-party data. Taryn Dominie, Head of Industry at Orange Apron Media, a division of The Home Depot, said translating shopping signals enables personalized media for diverse demographics, including Spanish-speaking customers and the next generation of Millennial and Gen Z homeowners. The partnership with Behr Paint Company has produced a 36% increase in purchase intent, Dominie noted. Andy Lopez, SVP & Head of Global Marketing at Behr Paint Company, said Orange Apron Media builds custom audience datasets to reach their professional buyers.

Ad tech integrations are also addressing interoperability friction and closing the gap between top-funnel engagement and bottom-funnel sales. Andrew Hotz, Director, Consumer Brands, Government & Entertainment at Google, discussed a first-of-its-kind integration allowing a retail partner to pull audience signals and apply them anywhere on the web, including YouTube. Google touches approximately 86% of all purchase journeys, Hotz noted, including through AI overviews in search now reaching 2 billion people worldwide. Christine Foster, SVP, Kroger Precision Marketing at The Kroger Co., confirmed this partnership lets grocery data inform campaigns at mass scale while measuring outcomes down to the SKU level. “If you know the cart, you know the heart,” said Christine Foster, SVP, Kroger Precision Marketing at The Kroger Co. Foster also noted that YouTube sees approximately 90 million hours of shopping video watched, which Google plans to connect directly to measurable sales outcomes, reducing waste across the ecosystem in the process.

3. Cultivating Emotional Connection and Loyalty

Track Keynote: Using Customer Data for Effective Merchandising

Effective merchandising is not about choosing between data and intuition, but about using each for what it does best. Anu Narayanan, President, Women’s & Home, Anthropologie, described Anthropologie’s approach as using customer and demand data to identify repeatable patterns, localized opportunities and emerging signals—while relying on merchant instinct to interpret what those signals could become. As she put it, “Data doesn’t tell you everything. It won’t tell you what’s next.”

That philosophy was evident in Anthropologie’s footwear expansion. While the financial profile of footwear alone argued for caution, broader data revealed that shoe shoppers were higher-value customers who drove stronger traffic and purchased across multiple categories. Combining basket, trade-area and whitespace analysis showed that expanding footwear in stores could lift the overall market rather than simply shift digital demand to physical retail—ultimately justifying footwear shops in more than 200 locations.

Localization has also become central to Anthropologie’s physical retail strategy, with assortments tailored by store and market based on category performance, digital demand and local customer behavior, allowing each location to feel distinct while remaining recognizably Anthropologie.

Anu Narayanan, President, Women’s & Home, Anthropologie
Source: Shoptalk Spring

 

Track Keynote: Visions for the Future of Marketing & Branding

As commerce becomes increasingly automated and AI-mediated, the value of human perspective, authenticity and community-driven insight is rising—not diminishing.

Steve Huffman, Co-Founder & CEO, Reddit, positioned Reddit as a trust layer in an internet ecosystem that is becoming more “sanitized” by AI-generated content, arguing that while LLMs are powerful, they tend to aggregate and simplify information, whereas real purchase decisions often require nuance, subjectivity and lived experience. This distinction is particularly relevant in commerce, where consumers are not just asking “what is the best product,” but “what is best for me,” a question that depends on context, preferences and multiple perspectives. Reddit’s growing role in discovery—evidenced by users appending “Reddit” to Google searches hundreds of times per second—highlights a broader behavioral shift toward seeking unfiltered, peer-driven validation alongside algorithmic recommendations. Huffman said, “AI can summarize the internet, but it can’t replace lived experience.”

Steve Huffman, Co-Founder & CEO, Reddit
Source: Shoptalk Spring