How Kroger, Giant Eagle and DoorDash are building their agentic AI assistants
AI was the hot topic of conversation at this year’s Groceryshop, as executives from Kroger, Giant Eagle, DoorDash and more took to the stage to share their philosophies on agentic technology.
What many companies are finding is that rather than relying on simple search tools or intrusive chatbots, they must deploy practical AI tools that bridge the gap between physical stores and online ordering. Much of the focus now is on training shoppers to naturally integrate agentic shopping into their busy routines.
Yael Cosset, the chief digital officer at Kroger, said the retailer’s big focus is on removing any friction between the time when a customer is inspired to buy an ingredient for a meal and actually going to order it. “The shopping assistant is helping close the gap that exists between inspiration and basket shopping,” he said.
But for many large retailers, the main challenge lies in integrating agentic tools into shopping journeys that are already complex.
Shweta Prabhu, vp of digital, e-commerce, MarTech, loyalty, data analytics and AI, enterprise systems at Giant Eagle, said the company’s focus is on using agentic AI to turn single-channel shoppers into omnichannel shoppers. That’s because Giant Eagle’s omnichannel shoppers tend to have a 30% higher lifetime value when compared to shoppers that buy from one single channel, whether in-store or online.
As such, Prabhu said every dollar that the retailer invests in making those experiences seamless for users “is one that will keep our most valuable customers coming back to us.” On its main grocery app, Giant Eagle now has an AI recipe generator, which helps customers scan meal ideas and seamlessly add ingredients to their cart.
Another focus for Giant Eagle as it invests more in AI is doing a better job of integrating fragmented teams on the backend operational side.
Heather Feather, Giant Eagle’s vp of digital, e-commerce and customer analytics, explained that one way the retailer began doing this last year was by combining the e-commerce operations team with the marketing and merchandising teams. “No store receives as much weekly traffic as our website and mobile app do,” Feather said. “I think that’s a great reframe for all of us and our partners as we think about investment and priorities.”
By working more closely together, the departments can better target potential crossover in-store and digital shoppers. And there may be more integrations to come in 2027 — earlier this year, Kroger announced its intention to acquire Giant Eagle, with the deal expected to close next year.
Building a naturally conversational shopping assistant
Kroger’s Cosset emphasized the need for grocers to build out a steady strategy to help shoppers get accustomed to using agentic grocery shopping.
“Customers don’t wake up and say, ‘Oh, I’m going to shop online today,’ or ‘I can’t wait to go into the store,’” Cosset said.
Instead, busy people have many different needs and occasions that retailers like Kroger have to address. “It’s on us to remove that friction and the burden of having to think differently about how they engage with us,” Cosset said. “And that’s where I think AI really is transformational.”
Cosset said Kroger spent the past 10 years building its digital and physical experiences together. For example, shoppers should be able to see and track their online orders and in-store hauls in one hub, with that data informing their future shopping journey.
The retailer recently launched its own AI shopping assistant, which Cosset said aims to solve universal grocery shopping problems, like busy parents needing to make last-minute grocery runs or figuring out what to make for dinner when hosting someone with multiple dietary constraints. The tool then instantly creates a Kroger basket with items that address all those requirements.
Shoppers can also add personal context to inform recommendations, like a photo of a handwritten shopping list or even their grandma’s favorite recipe.
“The investments we’re making right now are to actually continue to inject the personalization capability we’ve built over decades into that shopping assistant,” Cosset said.
With this contextual shopping experience, Kroger’s CPG partners also have the opportunity to get Kroger’s shoppers to engage with their brand, whether through retail media advertising or recipe content.
Identifying what people don’t want from agentic shopping
DoorDash co-founder Andy Fang said that, for delivery platforms, the challenge is learning what users want from AI-powered shopping tools as the company builds them out.
“Oftentimes when people think of using AI to order, they think of a chat window popping up,” said Fang.
The problem DoorDash found is that when people see a chat window pop up, they’re not sure what to talk to the bot about and end up ignoring it. “So a lot of work that we’ve done is not only to incorporate intelligence into the experience,” he said. “But also picking up on the trends of actual problems that people want AI to solve for them.”
Fang said most users, instead, appreciate simple, practical tools. For example, they like the ability to automatically prompt “reorder my usuals” during the browsing journey, instead of sifting through their order history page.
He said DoorDash is trying to go beyond populating a cart by building tools natively into the app that make it easier to modify real-time orders, such as finding similar SKUs or alternatives to out-of-stock items.
While retailers are getting a lot of real-time feedback on how customers use agentic AI, there are still many unknowns and lessons to learn.
On the big pressure retailers face to deploy AI technology, Cosset said, “Of course everybody wants to go fast [because] it’s a huge opportunity. But with speed comes risk, so speed with precision and direction is good.”