It’s an age-old story. Retailers are constantly seeking technology to solve their problems. Self-checkout was supposed to ease checkout lines. Amazon Go was supposed to make self-checkout obsolete. Neither transformed the store. But what has transformed? Major brands, including Target, Walmart, Ralph Lauren, Best Buy, Gap, and Dick’s Sporting Goods, have C-level AI executives. These brands are taking the future seriously. Conversational commerce and digital agents are the next wave, and these tools have interesting possibilities for grocery and mass retail.
AI shopping assistants may be different, not because the technology is brilliant but because it is becoming more meaningful in shoppers’ purchase decisions. AI doesn’t just speed up a transaction; a good AI assistant can influence what gets discovered, which products are purchased, what gets substituted for an out-of-stock, and can make suggestions based on a shoppers’ history or personal criteria. Alexa is the shopping assistant pioneer as a tool in Amazon’s forward-looking commerce models. There is a potential trust issue, however: Which brands receive a recommendation and why?
Are AI assistants trusted or merely transactional? And the answer is: They are emerging in the grocery sector as a useful tool, but who’s programming the product recommendations?
Gold Rush
Walmart CEO John Furner told investors on his Q1 earnings call in May that shoppers using the Sparky AI Shopping Assistant spend 35 percent more per order than shoppers who don’t use it. That’s an impressive number. Amazon has reported stronger conversion from AI-assisted shopping. Sensor Tower data reported by Adweek found that 44 percent of app sessions using Alexa for Shopping converted to a sale in July, compared with 34 percent a year earlier under the previous Rufus tool.
Kroger has been expanding its AI-enabled shopping and meal-planning capabilities through Google Cloud. The appeal is obvious: A shopper can begin with a meal intention like “I’m having a BBQ for ten people, one person is gluten free, one is vegan, and another is diabetic, I don’t want to spend more than $7 a person.” Click a button and end up with a cart shaped by inventory, price, loyalty data, dietary preferences, fulfillment options, and suggested add-ons. Cart Happy is a new AI assistant that lives on a browser and acts as a personal grocery shopping assistant, remembering a user’s regular purchases, automatically comparing prices, adding coupons across stores based on a shopper’s zip code. And then it recommends which store to buy which products in what order to build the lowest-cost cart and then moves the products to multiple retailers’ checkout carts. This tool should make every grocer and brand nervous!
It all sounds good, but are we looking at solutions that are ahead of shopper behavior? Dunnhumby’s Consumer Trends Tracker found that only 15 percent of U.S. shoppers use AI tools for grocery shopping. A higher basket size among early users may show that the technology is valuable, but it does not yet prove that a wider grocery audience wants an AI helper in their shopping experience.
Who Do You Trust?
There’s a harder question sitting underneath all of this, and most retailers would rather not answer it: Who actually programs the recommendation engine, and in whose interest? Is it a suggestion that shows up because it genuinely fits what I asked for, or because a vendor paid for that shelf space inside the algorithm, or because the retailer’s own margin math decided for me? Shoppers can’t tell the difference right now, and almost nobody in the industry is volunteering to explain it.
We’ve spent a decade asking the same question about search results, social feeds, and news algorithms, and grocery just caught up to it. The retailers and tech vendors building these assistants need to be honest and transparent about how the recommendation logic gets built, who gets to weight it, and whether a brand ever outranks what the customer actually asked for.
When an AI assistant recommends a pasta sauce, soup, yogurt, or snack, the shopper needs to understand whether that product was selected because it best fits the request, because it is available and competitively priced, because the retailer wants to improve margin, or because a brand paid for placement. If it’s all about the retailer making money, the shopper will find out and stop using the tool.
Trust is also an issue with retail media networks. These pay-to- play in-store platforms are trying to turn retailer-owned digital media into streaming advertising. The strategy has been substantial incremental revenue with a growing risk of customer rejection. There are too many sponsored listings, screens, banners, and prompts that is making the in-store and online shopping experience feel cluttered, impersonal, and distracting.
On the plus side, AI assistants offer a potential reset because they feel more personal and relevant. These agents are notorious for anticipating needs and seamlessly offering well-timed product recommendations along with the ability to order, purchase, and ship them. But success depends on trust and transparency and avoiding creeping out their owners.
Disclosure and Relevance
Kroger, Walmart, Amazon, and Albertsons are moving toward a model in which conversational commerce, retail media, and transaction data increasingly overlap. When done right, it’s a valuable asset for consumers. But retailers should establish a few basic rules by clearly identifying sponsored/paid recommendations when they appear within conversational responses. Retailers should resist the temptation to allow paid placements to override consumer needs, particularly in health, allergy, dietary, or budget-sensitive requests.
The industry has spent enough time learning that short-term ad yield can weaken the experience that makes that ad yield possible. AI assistants only raise the stakes, because a bad ad gets ignored while a bad recommendation feels like a betrayal. A shopper might shrug off a bad ad, but once an AI assistant feels like it’s been leading them through the shopping experience, they don’t come back. Retail media has become a powerful business because retailers own valuable purchase data and high-intent consumer moments, but when it starts to feel like we are being manipulated the retailer loses.
My Bottom Line
AI assistants could do what retail media networks promise but often fail to deliver: increased basket size without making the shopper feel relentlessly barraged by transactional ads. The macro-opportunity for any retailer is to connect a shopper to information that changes a purchase decision and delivers a better customer experience. If an AI assistant can do that, the tool is an asset, not a gimmick. Think like a shopper (or when you are shopping). Who or what do you trust? Are you the center of attention? Or a transaction waiting to happen?
