Deciding what to make for dinner every night is tiresome. Grocery stores have a solution: Just ask AI.
AI shopping assistants can now suggest recipes, recommend products and even build an entire grocery cart from a simple request. Walmart’s Sparky, for example, can turn a “What’s for dinner?” prompt into meal ideas and add the ingredients to a shopper’s order.
There’s just one catch for those trying to save money: People using these AI tools appear to be buying more. Albertsons, Walmart and Instacart have all reported larger baskets or higher sales among shoppers using their AI-powered shopping assistants.
Consumers are quickly getting comfortable with the technology. A recent survey found that 67% of shoppers have used an AI chatbot to help with shopping in the last three months. During last year’s Black Friday and Cyber Monday cycles, 68% of shoppers who used AI said they spent more than they planned, compared with 25% of those who didn’t use AI.
This is a correlation, not proof that AI caused overspending, but the pattern is worth paying attention to as grocery stores increasingly insert artificial intelligence into the shopping experience. The technology is supposed to make shopping quote-unquote “easier.” But could the convenience sneakily lead you to spend more than you planned?
The answer depends on what you ask the AI assistant to do.
The problem with making grocery shopping too easy
Grocery shopping involves a lot of pesky decisions. Shoppers have to figure out what to buy, compare brands and prices, find the ingredients and determine whether they’ll actually use everything they put in their cart.
“The old shopping model has plenty of friction points,” says Marty Bauer, an e-commerce and retail expert at Omnisend, a marketing automation platform. “A customer has to make a ton of decisions that can be quite exhausting.”
That’s where AI shopping assistants come in. They can take some of those choices off shoppers’ plates by suggesting meals, narrowing down product lists and even building carts based on special requests.
This technology can be particularly useful when shoppers are overwhelmed, according to Colleen Kirk, an associate professor at New York Institute of Technology who studies consumer behavior.
“When consumers shop on Amazon, they are confronted with a myriad of choices, features, colors, brands [and] reviews,” Kirk says. AI can reduce the cognitive burden by “filtering out less likely options and making recommendations based on a consumer’s needs, interests and prior shopping history,” she adds.
Those recommendations may carry more weight because AI assistants can seem surprisingly human. Generative AI agents are designed to be warm and casual, Kirk says, which can create a sense of social presence and increase consumers’ trust in their suggestions.
Indeed, there are already sign that AI-assisted shopping may lead to bigger grocery orders.
An Albertsons executive told the Wall Street Journal that shoppers using its AI-powered conversational search feature had 10% higher average order values than those using traditional search. For shoppers opting for a more comprehensive experience, which can help with recipes and dietary needs, average order values were 26% higher.
Walmart has reported a similar pattern. Shoppers using its AI assistant have baskets roughly 35% larger than those who don’t use it. Instacart also said orders placed with its newly launched AI assistant are, on average, bigger than its typical basket.
I tested some of the technology myself by asking Walmart’s Sparky to help me make tacos for a family of four. It offered two options, including a full ingredient list for $5.23 per serving and another for $6.70 per serving. When I asked whether there was a cheaper alternative, it immediately offered the “most budget-friendly” option for $3.66 per serving.
That small experiment illustrates how much the prompt a person writes can shape the shopping experience (and eventual grocery bill). Yes, the AI could help me spend less — but I had to explicitly tell it that was my goal.
The larger concern leading to overspending may not be one unnecessary purchase but a series of small, seemingly reasonable ones.
Put simply: An old-school shopping list has a natural stopping point. AI doesn’t.
Clay Cary, a senior trends analyst at CouponFollow, says the technology can subtly shift the question from “What do I need?” to “What else could I use?” An assistant helping someone plan taco night, for example, might recommend salsa in addition to the ingredients the shopper originally had in mind.
Oh, and guacamole. And don’t forget the drinks.
“Every single addition may sound perfectly reasonable on its own, but five or six ‘helpful’ additions will lead to a bigger bill than anticipated,” says Cary. “In this regard, personalization turns impulse purchasing into something almost natural.”
How to use AI to spend less on groceries
This doesn’t mean shoppers trying to save money need to avoid AI altogether. In fact, Kirk says the technology can help budget-conscious consumers, especially because they can use natural language to sort through a retailer’s offerings.
“If you have a budget in mind, tell the AI assistant,” Kirk says. Giving it specific information about what you’re looking for, such as price or brand, can narrow the choices.
Instead of simply asking an AI assistant to plan a dinner for a family of four, a shopper should give it constraints. For example, say something like “Plan five dinners for my family of four for under $75. I want to prioritize store brands and reuse ingredients across meals.”
The goal is to make the AI assistant work toward your priorities rather than simply letting it decide what belongs in the cart.
And be careful not to let AI fill your cart too quickly. Once a product is in your cart, it’s common to begin to feel a sense of ownership over it, Kirk says, making you less inclined to remove it.
The point isn’t to avoid grocery-store AI assistants. It’s to make sure you’re the one deciding which part of the shopping experience you want to optimize.
