That’s the connective thread running through the opening episodes of The SKU, a PYMNTS On Air series presented by Marqeta that spotlights the ideas and innovations reshaping retail. In conversations with Parbinder Dhariwal, VP and GM of CVS Media Exchange, Armand Wilson VP of categories and expansions of Whatnot, Josh Friedman, senior vice president of eCommerce and digital of Ulta Beauty, Dan Bennett, CMO of Furniture.com and Newtimes Group’s managing director Alex Angelchik, a common agenda emerges: use smarter technology and richer data to collapse friction, modernize margin and make every interaction feel more like a relationship and less like a transaction.
By design, The SKU sits at the intersection of the physical and digital economy. Each episode is a candid conversation about how executives are reimagining growth, loyalty and relevance in categories that range from pharmacy and beauty to home goods and luxury apparel. It is less about channel strategy in the narrow sense and more about the operating systems underneath: how live-commerce marketplaces monetize communities, how retailers become media businesses, how supply-chain specialists become brand owners, and how all of that changes what “shopping” looks like to consumers who move fluidly between social feeds, stores and screens.
AI: Plumbing, Not Pyrotechnics
AI is the most visible new tool in the kit, but the way these leaders talk about it sounds more like plumbing than pyrotechnics. At Whatnot, Wilson insists the stars are human sellers, not algorithms. AI is an “efficiency layer” that lets a seller hold product up to the camera and have a listing and auction created instantly while trust and safety checks run quietly in the background.
At Furniture.com, Bennett uses in-house AI to digest feeds from roughly 70 retailers and normalize them into a single, structured catalog — and is now building for a prompt-based search world where “I need a sectional for three kids and a dog” becomes a set of specific, shoppable options instead of a 15-hour browsing slog.
Ulta’s Friedman is experimenting with a virtual beauty advisor that uses large language models (LLMs) to answer routine and beauty questions conversationally, while AI-driven replenishment and recommendation engines anticipate guests’ needs in the background. In all three cases, AI is the invisible engine that removes friction so people — sellers, associates, guests — can spend more time engaging and less time hunting, typing or waiting.
Advertisement: Scroll to Continue
Data as the Connective Tissue
If AI is the engine, first-party data is the fuel.
CVS sits on 90 million addressable consumers across beauty, health and wellness, giving Dhariwal’s retail media network a uniquely granular view of behavior: the shopper who never switches lip gloss brands but happily experiments with facial cleansers, or the ExtraCare member whose in-store and digital journeys need to be orchestrated rather than bombarded.
That allows CVS to sell media that is rooted in real elasticity, not guesswork, and to connect the dots across email, app, offsite media, in-store screens and receipts.
Furniture.com’s normalized catalog lets the company understand demand at the level of finish, configuration and budget across partners, while content and engagement data from its growing library of articles and short-form video feeds back into SEO and merchandising. Ulta’s 45 million loyalty members provide both explicit and implicit signals that power its curated marketplace and its replenishment and personalization engines.
Whatnot watches its “and whatnot” category to spot emerging communities and verticals, identifies breakout sellers and formats, and tunes its rewards program using observed behavior, driving more multipurchase activity without relying on bluntforce discounting. Newtimes, meanwhile, uses real-time reads from a single Robert Talbott boutique — what sells, at what price, in which fabric and fit — to inform assortment and sourcing decisions across a global manufacturing network that spans dozens of countries.
The New Store Math
Those data and AI foundations exist to support, not supplant, the physical experience.
In beauty, Friedman is blunt: stores remain “paramount.” Ulta’s digital investments — from virtual try on to its invite-only marketplace and AI experiments — are designed to be the mortar around those bricks, ensuring that any piece of inventory can serve any guest and that inspiration from social can be turned into a seamless in-store or digital journey.
Furniture.com is a technology company “in the service of furniture,” but Bennett noted that 70% to 80% of large purchases are still made or initiated in-store. The site flags when a product is available locally so shoppers can go sit on the sofa they found online, sending better-qualified traffic back to partners instead of displacing them.
CVS is turning dwell time at the pharmacy counter into a media asset, using waiting area screens in roughly 2,000 stores to tell targeted product stories and then sending shoppers back through aisles on the way out. Newtimes chose to reopen Robert Talbott in the exact Madison Avenue space it had exited years before, deliberately using a flagship store as a brand lab and storytelling stage in a reboot that also includes wholesale and D2C eCommerce. And Whatnot has effectively rebuilt the feeling of wandering a favorite shop — the serendipity, the recognition, the regulars — in a purely digital environment where the goods are physical but the relationships are native to the stream.
The Customer as OS
In every case, the technology is in service of a very analog goal: making individual customers feel seen, understood and fairly treated.
Whatnot began as a single community of Funko Pop superfans; its continued growth depends on staying close enough to each community to know which categories and tools to build — and how much automation the culture will tolerate before it feels inauthentic.
At CVS, Dhariwal argued that retail media must fit “into the hand that fits into the glove of the retailer,” meaning that loyalty, merchandising, marketing and media have to work together so consumers experience one coherent CVS rather than a clutter of disconnected offers. Ulta’s ambition is to make channels irrelevant from the guest’s perspective; whether she starts with a viral TikTok, a marketplace search, a replenishment email or a stylist consultation, the experience should feel like a single relationship wrapped in the same loyalty benefits.
Furniture.com focuses relentlessly on confidence, recognizing that furniture is often a household’s third-largest purchase after a home and car. Bennett’s mandate is to strip redundancy and uncertainty out of the process so what should be a joyful, creative project doesn’t become a bandwidth-draining chore. Angelchik, for his part, is willing to accept that evolving a heritage menswear label toward a modern value proposition will cost some legacy followers if it means earning a new cohort that sees Robert Talbott as a fair trade between craft and cost.
For financial institutions, networks and fintechs, these stories are not just colorful case studies from the edge of commerce. They are blueprints for how value will be created — and where payments and embedded finance will need to show up — in an economy where shopping is increasingly a stream, a service or a story as much as a SKU. For issuers and acquirers, that means the real battle will be won in the orchestration layer: being the trusted rails that make these new, media-driven, data-rich commerce flows safe, seamless and measurable.
The SKU conversations make clear that the next era of commerce will be won not by those with the flashiest storefronts or the biggest tech budgets, but by those who quietly put AI and data to work behind the scenes while elevating, not erasing, the people and places that make shopping feel like something more than a transaction.
As Wilson put it, Whatnot’s mission — and, increasingly, retail’s — is simple: “We’re adding the personal element to the shopping experience that has been missing as eCommerce has just taken off.”
