The features, in order
Target has been shipping AI-powered shopping features on a steady cadence rather than announcing a single flagship product. Continue Shopping arrived in fall 2025, reconnecting customers with recently viewed products. Review Insights followed in June 2026, using AI to summarize product reviews so shoppers can gauge fit, quality, and common complaints without reading dozens of individual posts. Photo search launched in August, letting mobile app users search for products by uploading or snapping a picture. Buy Again rounds out the set, surfacing frequently purchased items and relevant deals based on a shopper's own purchase history.
None of these features are individually remarkable. Photo search and review summarization exist at other retailers, and purchase-history-based recommendations have been standard ecommerce infrastructure for a decade. What is notable is the cadence and the discipline behind it: four distinct capabilities shipped over roughly a year, each narrow enough to measure cleanly, rather than one large agentic shopping assistant bundled together and shipped as a press event. That sequencing also means engineering and product teams could learn from each release before committing to the next, instead of betting a full roadmap on a single unproven concept.
The metrics behind the features
Target says Review Insights has already driven higher conversion rates and increased cart additions, a direct funnel metric rather than a vague engagement claim. Buy Again has contributed to strong year-over-year growth in repeat purchases, the kind of number that ties cleanly to retention rather than novelty usage. These are the metrics a retail CTO should actually care about when evaluating whether an AI feature earned its build cost, and Target appears to be reporting them because it can, not because a PR team asked for softer language.
That matters because most retail AI rollouts this year have leaned on adoption numbers, users who tried a feature, sessions with an AI assistant open, rather than outcomes tied to revenue or retention. Adoption is the easiest number to make look good and the weakest signal of whether a feature is actually worth its engineering cost. Conversion lift and repeat-purchase growth are harder to fake and harder to dismiss, which is exactly why they are more useful to the executives who have to decide what gets funded next.
Sarah Travis's framing is the tell
Sarah Travis, Target's executive vice president and chief digital and revenue officer, described the strategy plainly: "Guests move naturally between our stores and digital channels, and we're using AI and personalization in purposeful ways to help them find what they need faster, discover new possibilities and shop with confidence." Notice what is absent from that statement. There is no mention of agents, autonomy, or AI making decisions on the customer's behalf. Every feature described is assistive: it helps a human shopper move faster through a process they are still controlling.
That framing is a deliberate contrast with the agentic commerce narrative dominating industry conversation this year, where AI increasingly makes purchase decisions with the shopper's authorization but limited moment-to-moment involvement. Target's bet is that most shoppers, most of the time, still want to see the product, read the summary, and click buy themselves. It is a more conservative technology position, and it is also the one with cleaner, more defensible metrics behind it right now.
The Q2 backdrop makes the timing look smart
This rollout lands alongside genuinely strong financial results. Target's net sales rose 5.3 percent year over year to 26.5 billion dollars in Q2, and net earnings roughly doubled to 1.9 billion dollars. That performance gives the AI feature rollout credibility it would not have during a weaker quarter, since Target can point to funnel metrics that sit inside a broader growth story rather than metrics offered as a distraction from underlying softness. Investors and analysts are more inclined to believe measured AI claims when the company is not also explaining away a miss.
It also buys Target patience for the incremental approach. A retailer posting soft numbers would face more pressure to announce something bigger and more attention-grabbing, a shopping agent, a generative styling tool, something with a press-friendly demo. Target's Q2 strength means it does not need the AI story to carry the earnings narrative on its own, which is precisely why the AI story here reads as more credible than most.
The holiday backdrop this is shipping into
Deloitte's holiday forecast projects overall retail sales growth of 4 to 4.8 percent this season, with ecommerce outpacing that at 7.5 to 8.4 percent growth, reaching roughly 316 to 319 billion dollars. Deloitte economist Akrur Barua attributes the ecommerce outperformance to consumers' ongoing use of digital tools to research, compare, and complete purchases. Separate Bain research found that 24 percent of holiday shoppers now plan to begin product discovery through AI tools like ChatGPT, Gemini, or Claude, a 17 percent increase over last year.
That is the environment Target's incremental features are shipping into. Photo search and review summarization are not exotic bets on a hypothetical AI-driven future. They are practical responses to a shopper behavior shift that is already measurably underway this holiday season, built on infrastructure Target could ship and iterate on in months rather than years. Retailers still deciding whether to build a full agentic assistant or ship narrower, measurable AI features should treat this rollout as a data point worth studying closely.
What this means for your own AI feature roadmap
The lesson here is not that agentic commerce is wrong or that Target has the definitively correct strategy. It is that shipping narrow, measurable AI features tied to funnel metrics is a lower-risk path to proving AI ROI than committing to a large agentic platform bet before the underlying use case is validated. Target can point to conversion and retention numbers today. Retailers who bet everything on a single flagship agent often cannot point to anything comparable for another year or two, because the scope of what they built makes clean measurement much harder.
If your organization is still debating whether to chase the big agentic commerce story or ship smaller AI capabilities against existing funnel metrics, Target's cadence is a useful reference case. Start narrow, instrument every feature against a business outcome you already track, and let the bigger platform bets wait until you have evidence that shoppers actually want AI making decisions for them rather than just helping them decide faster themselves.
