Target's SVP of Technology told Retail Dive that guests who build a wish list drive about 45% higher demand in the category. The more interesting detail is what's actually running underneath that number, and the team it takes to build it.
Target's back-to-school shopping experience is being powered, in part, by AI this year. The company told Retail Dive that customers who build wish lists drive roughly 45% higher demand in the category, and it's using AI to make that wish-list process, and back-to-school shopping more broadly, a more guided experience.
Target Senior Vice President of Technology Brad Thompson told Retail Dive that the retailer uses AI to generate product recommendations as shoppers start building a wish list, nudging them based on what similar guests have bought, what they've purchased before, and what a given school's supply list might call for.
What Target described to Retail Dive isn't a single AI feature. It's three, working together. The first is the wish-list recommendation engine above. The second is what Target calls its "next best action" widget, a feature that lives on the site year-round but gets recontextualized seasonally, during back-to-school, the suggested action might be starting a wish list. Thompson said Target uses AI to read the signals in a shopper's browse session and decide which single action to surface next, rather than showing every visitor the same static prompt.
The third, and the most forward-looking, is content personalization at scale. Target already serves different shoppers different homepage experiences based on what it knows about them, and Thompson described a near-term future where that extends to literally configuring the homepage, a product detail page, or a product listing page differently for each visitor, based on AI's read of that individual.
| AI Feature Target Described | What It Actually Requires to Build |
|---|---|
| Wish-list product recommendations | Behavioral modeling that connects past purchases, browse signals, and catalog data into an individual suggestion, not a generic best-seller list |
| "Next best action" widget | Real-time processing of in-session signals, translated into one specific on-site prompt, reusable and recontextualized across seasons |
| Page-level content personalization | Infrastructure that can dynamically assemble home, PDP, and PLP layouts per visitor, built as a reusable system rather than a one-off test |
We've written before about the gap between AI adoption and AI scaling in retail: roughly 89% of retailers have adopted AI in some form, while only about 7% have actually scaled it into how the business runs, a gap we covered in our piece on hiring a Head of AI Transformation for retail. Target's back-to-school push is a working example of what that other 7% looks like in practice.
It isn't one AI feature bolted onto an existing site. It's three connected systems, a recommendation engine, a reusable decisioning widget, and personalization infrastructure meant to eventually run at the page-layout level, all sitting under a single technology leader who treats a predictable seasonal traffic surge as a deliberate testing window rather than just a merchandising calendar event.
Adopting AI usually means one feature, tested once. Scaling it means building the underlying decisioning and personalization infrastructure once, then reusing it across every seasonal moment that follows. Target's back-to-school push is the second thing, not the first.
The stakes behind that testing window are real. Spending for K-12 students is projected to reach $43.3 billion this year, while college student spending could reach $103.5 billion, according to the National Retail Federation and Prosper Insights & Analytics. That volume is exactly why Thompson's team treats the season deliberately: "such a great seasonal moment for us," he told Retail Dive, is also a recurring, high-traffic window his team can use to validate new features quickly, before rolling them into the rest of the year.
That's a meaningfully different posture than treating a seasonal spike as purely a marketing and merchandising event. Target's technology organization is using it as a scheduled experimentation window, features get built ahead of the surge specifically so they can be A/B tested against real volume, then kept, tuned, or dropped based on what the data shows.
Most retailers reading this don't have Target's traffic volume, and don't need it to apply the same approach. The lesson isn't about scale, it's about treating personalization as an ongoing product surface with an explicit owner, rather than a project that ships once and sits static until someone remembers to revisit it.
That requires two things most retailers haven't explicitly staffed. The first is a data science or applied personalization hire who can turn raw behavioral signals into a single, real-time recommendation, the kind of role we cover in our Analytics & Data Science practice area. The second is a technology and site engineering function willing to build that recommendation as reusable infrastructure rather than a one-off homepage banner test, closely related to the Omnichannel Customer Experience hiring gap we've written about, since a next-best-action prompt is, at its core, a customer experience decision made by a model instead of a merchandiser.
Retailers that have already read our piece on the great unbundling of the omnichannel manager role will recognize the pattern here too. Personalization at the level Target is describing doesn't live cleanly inside marketing, technology, or CX. It needs explicit ownership at the seam between them, or it stays a one-time feature instead of the reusable system Target is building.
Target didn't get to three connected AI systems by declaring an "AI initiative." It got there by staffing the specific pairing of behavioral modeling and reusable site infrastructure, then giving that pairing a recurring, high-stakes season to prove itself against. Most retailers have the season already. Very few have staffed the pairing.
Target's SVP of Technology, Brad Thompson, told Retail Dive the retailer is using AI in three specific ways this back-to-school season: recommending products as shoppers build wish lists based on what similar guests bought and what a given school's supply list might require, a "next best action" widget that prompts a specific on-site action contextualized to the season, and early-stage AI-driven content personalization aimed at eventually configuring homepage, product detail, and product listing pages differently for each individual visitor.
It's a feature that lives on Target's site year-round and prompts a shopper toward a specific action, but the prompt itself changes seasonally. During back-to-school, that might mean nudging a guest to start a wish list. Target uses AI to process the signals in a shopper's browse session and decide which single action to recommend next, rather than showing every shopper the same generic prompt regardless of what they're actually doing on the site.
No, it illustrates it. We've written before that roughly 89% of retailers have adopted AI in some form, while only about 7% have actually scaled it into their operations. Target's back-to-school push is a working example of what that other 7% looks like in practice: not one AI feature bolted onto the site, but three connected systems, wish-list recommendations, a reusable next-best-action widget, and personalization infrastructure meant to eventually run at the page-layout level, all under a single technology leader treating a seasonal traffic surge as a deliberate testing window.
It typically takes a data science or applied personalization hire who can turn raw behavioral signals, browse history, past purchases, session activity, into a single ranked recommendation in real time, paired with a technology or site engineering team that can ship that recommendation as reusable, seasonally reconfigurable infrastructure rather than a one-off feature. Very few retailers have both halves of that pairing explicitly staffed, which is a large part of why AI adoption and AI scaling remain such different numbers.
No. The specific technology stack will look different at a smaller retailer, but the underlying approach doesn't require Target's size: pick a predictable seasonal traffic surge, decide in advance what new personalization feature you want tested during it, and make sure someone owns turning behavioral data into an actual on-site recommendation rather than a report nobody acts on. What Target's scale actually buys it is traffic volume large enough to validate a feature quickly, not permission to attempt the approach in the first place.