What's shaping the e-commerce landscape?

Target’s back-to-school shopping experience is being powered, in part, by AI this year.

The mass retailer found that customers who build wish lists drive about 45% higher demand in the category, the company told Retail Dive. The company’s using AI to help make the wish list process — and back-to-school shopping more broadly — an improved experience.

“We will use AI to make product recommendations as people are starting to create their wish list,” Target Senior Vice President of Technology Brad Thompson told Retail Dive. “So rather than them having to think about all of the things that they need, we will start to nudge them and prompt them based on what we know about other guests that have bought, or things that they've bought in the past, or what we might know about that school and the supplies that the teachers might need, and it's making a big difference.”

Target also has a feature called its "next best action widget," where part of the site will prompt a customer to take a specific action. The widget is regularly on the site, but it’s contextualized seasonally.

For the back-to-school period, that suggested action may be something like recommending a user create their wish list, Thompson said.

“We use AI to sort through all of the signals in the browse session for that guest to recommend the next best action,” the executive added.

The retailer in June debuted a slew of offerings for the back-to-school shopping season, including a deal on its Target Circle 360 paid membership, school uniforms from one of its private label brands, a limited-edition collab with LoveShackFancy and more.

Back-to-school shopping is an important seasonal event for retailers, and spending this year could reach record highs. Spending for K-12 students is expected to reach $43.3 billion this year, while spending for college students could reach $103.5 billion, according to an annual survey by the National Retail Federation and Prosper Insights & Analytics.

“It's such a great seasonal moment for us because we get a big surge in traffic, and so we try to have new features ready for the site,” Thompson said, adding that personalization through technology is one way Target is working to improve customer experience.

The back-to-school season allows his team to A/B test some new ideas. For example, Target is at the point where different people will get differing homepage experiences based on what the retailer knows about them — a concept that could be scaled with the help of AI, according to Thompson.

“We're starting to use AI to do content personalization at scale, and eventually we'll be in a place where we'll be able to literally do sort of one-on-one hyper personalized page layout,” he noted. “Depending on whether it's the home page, whether it's a product detail page or product listing page, we’ll be able to configure the site dynamically to appeal to you based on AI's recommendations.”

Walmart’s recent announcement that it has started delivering groceries in a half hour in dozens of markets might look like just another move by the retailer to keep up with Amazon, which said it was expanding its own 30-minute grocery delivery service.

Look beneath the surface, however, and the race between the two retail giants to satisfy consumers' desire for speed exposes an emerging trend that could eventually have profound implications across the grocery sector, according to industry analysts.

Amazon and Walmart have long posed a threat to brick-and-mortar grocers online because of their vast reach and deep pockets. But as they build out their 30-minute offerings, they are taking away perhaps the top benefit grocery chains provide customers — convenience — said David Bishop, partner at grocery e-commerce consultancy Brick Meets Click.

Rapid delivery is “undercutting one of the core strategic advantages that regional grocers or supermarkets have historically enjoyed in the physical world, which is proximity to the customer,” Bishop said.

Amazon’s 30-minute delivery service, Amazon Now, already serves several major markets, including Atlanta, the Dallas-Fort Worth area, Philadelphia and Seattle. The company plans to bring the service to dozens more cities by the end of 2026.

Walmart, meanwhile, currently offers its 30-minute service in 33 markets, noting that its network of stores gives it “significant opportunity to continue expanding delivery capabilities to more communities over time.” While Walmart formally announced the 30-minute service on May 28, it has been fulfilling orders that fast for more than a year, according to the retailer.

In another example of how Walmart is flexing its delivery muscles, the retailer announced that it would begin providing 30-minute restaurant delivery, starting with Subway locations in select stores.

Bishop pointed out that Amazon and Walmart pose an especially potent threat because they can quickly deliver basic goods that people are highly familiar with and buy often.

“If Amazon is able to demonstrate to households the ability to consistently and reliably deliver a quality product that’s a banana, to say the least, that’s concerning, because that has traditionally been one of the main trip drivers for supermarkets,” he said.

Michael Infranco, assistant vice president of RetailStat, observed that the recent steps by Walmart and Amazon to make grocery delivery more accessible are key reasons why traditional retailers are working with third-party providers to quickly scale up their e-commerce capabilities.

“They’re definitely putting pressure on [grocers], and you would expect them to keep running with that as long as they can,” Infranco said. “Anything that they can do to get a leg up on somebody else, they’re going to do via their size.”

In an example of how supermarket operators have looked outside for assistance in building up their delivery services, Kroger over the past year stepped up its partnerships with providers including Instacart, Uber and DoorDash. Kroger also works with those companies to provide 30-minute convenience delivery.

Albertsons and Ahold Delhaize have also looked to third-party providers to help them establish 30-minute delivery services.

Bishop noted that with its 30-minute service, Walmart is putting its stake in the ground as a “quick-trip” destination — a moniker the retailer has never had before.

Amazon, meanwhile, is gaining momentum by pivoting to better meet shopper expectations by putting groceries into the same category as its overall e-commerce business, Bishop said.

“They’ve realized, ’Hey, you know what? We’re creating friction, so let’s bring the products to the customer and put them where the customers are shopping, instead of putting it in the Whole Foods tab or the grocery tab.’ It’s now in the same-day tab or the Prime Now tab,” he said.

What’s in store for Amazon and Walmart in high-speed grocery delivery?

Bishop said he thinks that while Amazon’s dominant online presence gives it a powerful lever to expand its grocery business, its relatively small brick-and-mortar presence is a liability that it eventually will need to address.

The company’s decision in January to shutter its entire fleet of Amazon Fresh supermarkets left it with only Whole Foods Market as a foothold in the physical grocery store space.

On the other hand, Amazon’s ability to bring fresh foods into its enormous online fulfillment operations was a “seminal moment” for Amazon because it allowed the company to build momentum in the grocery space on turf it knows best, Bishop said.

“What’s clear is now, after 20 years of learning, Amazon has identified and aligned with its core shoppers currently, and that’s going to drive strong growth for them over the next year or two,” he said. “But eventually they’ve got to figure out the physical in-store presence, because without that they’re missing the pick-up segment.”

Pickup also presents an opportunity for Walmart to improve, Bishop added, noting that the company faces strong competition from Target — another mass retailer with a large physical store presence — in the pickup space. That could change if Walmart is able to speed up its pickup operations, including its efficiency in getting orders to people when they show up to collect them, he said.

“That’s helped Target compete against Walmart, because it has a point of competitive differentiation, which is the speed of its cycle times for drive up and its speed of service. It's still gold standard,” Bishop said.

Infranco said Amazon’s massive scale in e-commerce doesn’t guarantee it a leading spot in the 30-minute grocery-delivery space.

“Just because Amazon says they’re going to do it doesn’t mean they’re going to do it well or do it forever,” he said. “Grocery has been very hard for Amazon … they’ve been in that game for a while, and it seems like they’ve had just as many setbacks as they had victories.”

Infranco pointed to Amazon’s purchase of Whole Foods almost a decade ago as a development that didn’t turn out to be as disruptive as many in the grocery industry had anticipated.

“Amazon goes and buys Whole Foods, and everybody’s saying, ’Oh my god, the grocery business has changed forever.’ Well, you’re almost 10 years later [and] from the Amazon owning Whole Foods standpoint, the sector hasn’t changed all that much,” he said.

Sponsored content

By Mirakl

Reaching new customers keeps getting more expensive. Paid search and social cost more to reach the same audiences, while brands raise budgets to produce the same results.

There is another way to grow: selling through modern retail marketplaces.

That opportunity looks nothing like it did 10 years ago, when “marketplace” meant Amazon. Today, retailers such as Nordstrom, Macy’s, Bloomingdale’s, Best Buy, Walmart and Target run curated marketplaces, alongside hundreds of category specialists and regional players. These are stores customers already know and trust, looking for brands to add to their assortments.

The shift is underway. Sellers joining marketplaces grew 26% year over year, while marketplaces grew 4.5 times faster than traditional eCommerce, according to Mirakl Connect’s 2026 Seller Report.

Marketplaces are more than another ad channel

On social media, brands pay to put products in front of people who may not be shopping. On a marketplace, shoppers are already browsing a trusted retailer and looking to buy. Your products appear closer to checkout, with the retailer’s traffic and credibility working in your favor. Many marketplaces also offer advertising networks.

Siia Cosmetics initially saw marketplaces as an online discount rack. “Marketplaces are actually for very serious eCommerce now,” said Jessica Ahn, who runs operations for the K-beauty brand. The company found that curated retail partnerships strengthened its brand. After performing well on Macy’s, Siia was invited to join Nordstrom, and Macy’s introduced the brand to Bloomingdale’s.

You stay in control, without building a second business

Selling through a marketplace does not mean giving up control or creating a separate operation for every retailer. Brands choose where to sell, which products to list, how to price and promote them, and how they appear to shoppers. Unlike wholesale, where the retailer controls merchandising and inventory decisions, marketplace selling lets brands use a retailer’s reach while retaining more control over the customer experience.

The operational work can run through the systems brands already use. Tools such as Mirakl Connect’s Catalog Transformer reshape product data to meet each retailer’s requirements, reducing catalog work. For Siia Cosmetics, a new product that once took days to prepare now takes about two hours. Jessica manages the brand’s marketplace channels herself, overseeing orders, customer service and inventory in one place.

That approach can make growth more efficient. Rokka & Rolla expanded to five marketplaces and grew GMV by 80% in two years without adding operational costs. “We can connect with major retailers like Macy’s and Nordstrom much quicker and more efficiently,” said Ally Cheong, the brand’s business development manager.

Marketplaces can expand, not cannibalize, direct sales

A marketplace sale may not provide the same customer data as a sale on your own site. But it can introduce your brand to shoppers who might never have discovered it. Some then visit your site or stores to explore the full brand experience.

Ana Luisa saw its marketplace revenue grow from seven figures to eight in 18 months, alongside an increase in direct sales. “People came into our retail stores saying, ‘I discovered Ana Luisa on Macy’s, Bloomingdale’s or Nordstrom, and then wanted the full brand experience,’” said CEO and co-founder Adam Bohbot.

The next move depends on where you are today

If you do not sell through marketplaces yet, start with two or three retailers whose shoppers match your target customers. If you already sell on one or two, use sales data to choose the next marketplace strategically. If you are active across several, consider expanding into new regions or deepening the assortment where you already perform well.

The right marketplace fit matters more than being everywhere. As Angel Cerdá’s Diego Sierra put it, “We need to know if the target of the marketplace matches our products and our selling prices.”

Marketplaces are no longer a last-resort channel. They help brands reach high-intent shoppers through trusted retailers while keeping control of their brand and operations. The brands that build those relationships now will be better positioned for the next phase of commerce.

Ready to find the right marketplaces for your brand? Book a demo to see how Mirakl Connect can help you activate new growth opportunities.

Omnichannel retail has evolved to include artificial intelligence and agentic agents, but stores still need to focus on customer experiences and brand loyalty in order to drive traffic, executives from Ulta, Stitch Fix and other companies said during a June panel at the CommerceNext Growth Show in New York.

In order to meet consumers where they are, retailers need to understand how consumers are shopping, said Kimberly Shenk, CEO of marketing technology firm Novi, during a panel covering the future of omnichannel retail. “Most consumers are shifting discovery to agents, agentic platforms, ChatGPT, Gemini,” she said, adding that as a result, brand identity has to be consistent across every touchpoint.

Omnichannel isn’t only about in-store and online anymore, Josh Friedman, senior vice president of e-commerce and digital at Ulta Beauty, added during the panel. “We are seeing a lot more use of our app in the store, especially by our guests, maybe by our associates,” he said. “Taking care of those experiences is important.”

When companies better understand how AI operates within the omnichannel shopping experience, they’re better able to position themselves for success, said Shenk.

“AI recommends products as a solution to a problem that a consumer is asking,” Shenk said. “It's not recommending a brand name.”

As a result, brands need to position themselves within the fan-out queries of large language models, or LLMs, which do their own research to answer consumer questions. “[AI] doesn't actually take the user prompt and use that verbatim. It takes a user prompt and then triggers its own set of thousands of queries.”

However, search results don’t always lead directly to shopping carts, Shenk said. “A lot of this isn't leading to a click, because then, of course, the consumer loves their Ulta beauty points in their program,” she said. “They want to go back there, and so that loyalty component is still important.”

Friedman said that Ulta has 47 million loyalty program members, and that “95% of our sales go through the loyalty program.” Connecting that customer experience to AI helped the company “develop better algorithms, generate some better content, and more content,” Friedman said. The company’s recently launched Ulta AI has created a more integrated omnichannel shopping experience, he added.

Panel moderator Jon Kosoff, chief digital officer of Boot Barn, also talked about integrating his company’s loyalty program with its online, in-store and in-app experiences. He said that 30% of Boot Barn’s orders are currently picked up in store and 40% are shipped from stores, but that across all channels, loyalty was the most durable connective tissue.

“We just relaunched our app in April,” Kosoff said, adding, “Loyalty is key to that piece, and we just relaunched our tier-based loyalty program.”

Meanwhile, Debbie Woloshin, chief marketing officer at Stitch Fix, said that technology was a means to customer engagement.

“We do have an app, and about 60% of our clients have downloaded the app and use it regularly,” Woloshin said during the panel. “But what I would say, to take a step back from a loyalty perspective, is that we're basically in the relationship business.”

She said that her company’s clients often shared “very intimate information with us,” and using technology tools in combination with human advisers was the key to helping build and maintain those relationships.

“So, for instance, when someone is pregnant, we’re sometimes the first to know,” Woloshin said, adding, “There's a lot in play beyond just sort of a rewards feature to really build and deepen the personalization.”

Part of creating a durable customer relationship also comes from creating an omnichannel experience that feels consistent across every platform, said panelist Mandeep Bhatia, senior vice president of global digital product and omnichannel innovation at Tapestry.

“Earlier, it was hard to scale that store concierge experience, that omnichannel experience, online,” Bhatia said. “Now with agentic AI, we can. We talked to hundreds of store associates and said, ‘What do the customers ask when they come to stores? How do they talk to you?’”  

Bhatia said Tapestry trained its bots on those conversations in order to create an experience that was unique to the brand. “We want them to have the same experience that they have in stores, so that's how we think of the omnichannel architecture, not just from a technology perspective, but also from brand perspective,” he said.

Omnichannel is about consumers, not computers, Bhatia added.

“Don't start with technology first,” Bhatia said. “I mean, just because AI is there doesn't mean that we have to start creating those things. It has to be rooted in the understanding of customers and the understanding of associates.” He added that technology is creating an experience that’s “either painkillers, vitamins or chocolates, right? So you have to figure out how you balance that, but [that] will only come from understanding what a customer needs.”

TikTok has drawn droves of marketers for its ability to turn products of all stripes, from Stanley tumblers to cranberry juice, into viral sensations. Since the social platform launched its Shop e-commerce marketplace in the U.S. a little over two years ago, brands have gotten better at quantifying just how much video crazes are translating into sales.

TikTok Shop’s stature is growing enough to influence how retailers think about product mix, demand forecasting and content strategy, according to speakers at the National Retail Federation’s Big Show.   

“It’s a place for us to learn what’s working versus not,” said Richard Cox, chief merchandising officer of the Gen Z-focused retailer Pacsun, during a panel at the trade conference. 

At the same time, Shop has gotten more crowded and competitive. It acts as a “bonafide retailer,” one NRF speaker said, hosting its own Prime Day-like deals bonanzas called Super Brand Days. Shop is robust enough at this point that TikTok no longer offers brands some of the incentives it once did to use the feature. A fresh level of maturity comes as the retail industry is beset by challenges related to tariffs and pullbacks in discretionary spending.

“Our main operational challenge is around profitability,” said Jenna Manula Linares, vice president of digital marketing at Tarte Cosmetics, during the panel. “As their platform has scaled, they started pulling back in what they’d been funding for us.

“We also know that the TikTok customer is value-driven,” added Linares. “So now we are at this intersection where we’re trying to find the balance between how much value [we can] offer a customer while still being conscious of our bottom line.”

Sales follow virality

TikTok Shop increasingly seems like one of the first major success stories for social commerce in the U.S. The marketplace accounted for about one-fifth of the social commerce segment in 2025 and is forecast by eMarketer to exceed $20 billion in sales this year.

Pacsun’s first pop on TikTok Shop came as something of a surprise. Around Black Friday in 2023, a smaller influencer posted a video about the retailer’s Casey jeans, a low-rise, baggy cut of denim. The availability of the item on TikTok Shop dovetailing with a key holiday sales window resulted in 11,000 pairs sold on Black Friday alone, along with a long tail of popularity.  

“We’ve sold over 100,000 pairs of that jean. In terms of halo effect, it’s helped our entire denim business,” said Cox. 

Other brands on the panel shared similar case studies that speak to how Shop can link buzzy content to business results, sometimes in a chaotic fashion. Last year, Tarte — among the early adopters of Shop — noticed that creators were participating in a strange trend: They would draw under their eyes with a permanent black Sharpie and then cover up the markings with the brand’s CC under-eye color corrector, a testament to the concealer’s efficacy. Tarte has sold nearly 600,000 units of the product on TikTok Shop in the U.S., according to Linares, while noticing stronger demand in international markets through direct-to-consumer channels.    

That said, Shop requires a different approach than a traditional DTC or e-commerce site. Tarte promotes a smaller assortment on Shop because the brand “can’t control the algorithm,” Linares explained. Speakers noted that TikTok’s unpredictable nature is disrupting some of the fundamentals of retail, increasing a reliance on social listening, media mix modeling and “analytics horsepower.”

“It kind of turns everything we all know about demand planning on its head,” said Feliz Papich, senior vice president of digital technology, experience and insights at Crocs. 

Go with the flow

By that same token, marketers are adjusting some of their brand content to feed the Shop pipeline. Tarte, for instance, has introduced a mascot named Shapey, based on its shape-tape concealer, to attract viewers organically to its profile. 

“If you come to our page, you’re actually not going to see a ton of tutorials or before and afters,” said Linares. “We’re doing a lot around humor.”

The discussion also touched on co-creation, the idea of enlisting everyday customers to create content on behalf of a brand. Some retailers may not like relinquishing that degree of control but ultra-polished ads don’t perform as well in TikTok environments, according to Cox. Additionally, high-performing organic content is becoming valuable fodder to convert into paid media or material for other social platforms.

A willingness to roll the dice on such marketing experiments is just one piece of the Wild West mindset needed to succeed on TikTok, an app that is still in the process of figuring out its future in the U.S. following a ban threat.

“TikTok wants to partner with brands that are willing to take smart risks and to move fast. They’re not going to read your legal red lines,” said Linares. “Just say yes and go with the flow.”

Costco’s digital experience enhancements were a significant sales driver for the second consecutive quarter,  CFO Gary Millerchip said. Personalized product recommendations carousels contributed just under $500 million in digital sales, following $470 million the quarter prior. 

The digital carousels drove conversion rates three times higher than Costco’s typical conversion rate, according to Millerchip.

“Delivering a more personalized experience for our members is a key focus, and we continue to make progress in this area,” Millerchip said.

Costco’s digital growth was a highlight for the warehouse club, and the leaders have put an emphasis on increasing membership renewals from digitally focused customers.

Traffic to Costco’s website and app was up 37% during the third quarter, according to Millerchip. Multiple product categories saw double-digit digital growth, including pharmacy, home furnishings and housewares. 

However, members who sign up online renew at a slightly lower rate than members who join at a physical location, according to Millerchip. As a result, the growing share of digital-first members has applied downward pressure on Costco’s overall renewal rate.

Costco responded by implementing targeted digital communication and retention strategies aimed at these customers, Millerchip said. In the third quarter, this effort “more than offset the negative impact from this mix change in our membership base.”

The company is investing in multiple forms of technology to continue improving its experience both online and in-store.

AI “has the potential to be a significant sales opportunity for Costco,” Millerchip said. Volume from AI search remains low, but the retailer saw triple-digit AI search traffic growth in the third quarter, and this activity has the highest conversion rate of all website traffic.

The retailer’s other recent improvements include quick access to digital membership cards in the Costco app and improvements to the digital wallet, according to President and CEO Ron Vachris.

“In digital, we're making meaningful strides to deliver a more seamless and convenient experience for our members across the warehouse and online,” Vachris said.

In the U.S., e-commerce sales will reach $1.8 trillion by 2030, according to a new report from Forrester. 

Though U.S. e-commerce sales will comprise 29% of all retail sales, the remaining 71%, about $4.4 trillion, will come from stores, according to Forrester’s projections. 

Forrester attributes e-commerce sales growth to several trends, including more Gen Zers entering the workforce, improvements in logistics and order fulfillment, and the rise of agentic commerce. However, consumers will continue to shop in stores, because physical retail offers tangible experiences, in-person comparison shopping, social interaction and personalized assistance, the report said. 

Though online shopping is on the rise, brick-and-mortar stores have become critical not only for transactions, but also for creating unique in-person experiences and building loyalty. In the last 15 years, the value of in-store sales declined just once, in 2020 amid pandemic lockdowns, according to research last spring from Colliers. 

Still, online shopping remains a key channel for retailers, particularly during the holiday season. Between Nov. 1 and Dec. 31 last year, U.S. consumers spent $257.8 billion online, up 6.8% year over year, according to Adobe Analytics. Furthermore, traffic directed to retailers via generative AI tools skyrocketed over 693% year over year, Adobe found.

While Forrester projects that physical stores will continue to dominate retail, other research finds that the share of people who shop online every day has dropped from 21% to 9%, according to a recent Salsify survey.

With the exception of automotive and gas sales, total retail sales in the U.S. are forecast to grow from $5.2 trillion in 2025 to $6.2 trillion by 2030.

The promise of AI technology that is often repeated is it can transform how CX teams scale their operations, make employees more productive and open up access to data on unprecedented levels.

It doesn’t always work out this way, according to Brian Cantor, managing director at CMP. The benefits vendors promised from their AI solutions can fall short in unexpected ways or require more investment than initially expected.

“AI's supposed to augment everything,” Cantor said during a panel at Customer Contact Week Las Vegas last week. “Well, it can, but only if we do it the right way. And, unfortunately, sometimes we're being steered in the wrong direction.”

Companies can protect themselves from disappointment with the proper due diligence, according to experts. Leaders need to keep themselves from trying to implement too many capabilities, understand the cost of new technology, and invest with customers in mind to find solutions and vendors that fit their needs.

Sister publication CX Dive listened to leaders at CCW Las Vegas as they discussed the challenges of and strategies for cutting through AI hype. Here is a look at what four of them had to say.

Don’t get blinded by potential

AI technology is flexible, with applications that span seemingly every job at the contact center, according to Cantor. However, there can be too much of a good thing.

Cantor says CX teams can easily get caught up in the hype of individual capabilities without thinking about the long-term implications extra functions can create. At a certain point new tools can go from saving time to adding frustration.

“We have so many different things that functionally make sense as features, but ultimately, if adding a feature creates three more steps for everyone else or creates a new source of friction, that is something we have to start taking into account,” Cantor said. “I think we really need to understand the real ramifications of the real costs with each investment.”

The problem can be pronounced when AI tools fail to produce appreciable results, according to Cantor. The more time wasted on technology that doesn’t advance the CX team’s goals, the less flexibility the team may have in the future.

“You're not just going to be able to keep convincing your budget holders to keep trying a new AI platform until we find one that works,” Cantor said. “Each one puts your credibility on the line, you get your team's buy-in on the line. That's going to cause more inertia, more desire to stick with what you have, even if what you have ends up creating more friction.”

Scalability and cost

When a company rolls out a new technology, it can be easy to get excited by the initial wins — which can blind you to future scalability and cost concerns until it’s too late, according to Marius Maree, senior vice president of consumer operations at UnitedHealth Group.

A lot of companies don't ask about how well technology will scale beyond the initial pilot or how much it will cost to bring a project fully online before jumping into a new partnership, Maree says. However, it’s important for both parties to agree on what the long-term investment will look like.

“It's a tough conversation, but it's a partnership, so we try to meet each other in the middle,” Maree said during a panel. “I think that that's important. It's a commitment, so where do we draw the line?” 

That conversation may be hard, but it can be even worse when teams move ahead without accounting for the full cost, according to Maree. In that case, both sides need to hold a meeting after the cost has already exploded beyond the initial agreement.

While it can be intimidating, getting ahead of the problem usually isn’t an issue, according to Maree. Every partner UnitedHeath Group has worked with has been open to finding a compromise on scale and cost that works for both sides after the question was raised. 

Look beyond the demo to long-term considerations

It’s not hard to wow potential partners with a great initial pitch, according to Anant Singh, chief sales officer at Sanas, an AI real-time speech platform. The real work comes over the long term.

“Everyone does a great demo,” Singh said during a panel. “I don't think anyone has a bad demo or a bad meeting. What matters is longevity, repetition, consistency. You have to show up every time for months or years.”

The real work in a good relationship starts when a vendor earns its client’s trust, according to Singh. That’s not the time to celebrate — that’s the time to solidify the partnership by showing up every day and ensuring the client gets the quality service they were expecting. 

This is especially true for large, complex companies where tech investments can take years to pay off, Singh says. There, finding a partner that will remain reliable for years and can support the company’s overarching vision is more important than having the hottest technology of the moment.

“It's not about what's the coolest today — tomorrow that’s going to be something else,” Singh said. 

Investments start with the customer

The rapid advancement of AI has coincided with a rapid advancement of customer expectations, according to Melissa Solis, CEO of Inbenta, an AI solution provider. Solutions that don’t line up with their demands are doomed to failure.

“Customers are getting really smart, and they know what a good experience looks like, so they're demanding it,” Solis told CX Dive. “It's driving companies to have to relook, even if they have a solution, and say, ‘OK, is this the right solution?’”

In some cases, an insufficiently advanced chatbot can be a problem, according to Solis. ChatGPT and other major players have primed customers to expect human-level conversations with AI, and if a solution fails to deliver on that capability, some callers will immediately ask for a live agent — creating frustration for customer and business alike.

In other cases, companies need to slow roll their AI deployment and introduce it with care, according to Solis. For example, a business with an audience that’s largely 65 and older may find that its customers aren’t as ready to embrace AI agents as the general population.

Businesses should work with solution vendors to understand how businesses with similar customer bases meet their particular needs, including how to create an experience customers want to interact with that maintains the right level of human touch, according to Solis.

The lion’s share of e-commerce retailers have invested in agentic commerce in some form, most often to boost revenue growth and customer experience, according to a LogicBroker survey of over 600 enterprise e-commerce leaders. Nearly 96% have invested in AI capabilities.

Notably, the top applications are customer-facing, including Al-powered product discovery, Al chatbots, and personalized recommendations. 

Nearly half — 47% — plan to invest $1 million or more into agentic commerce in the next 12 months.

The days of businesses only deploying AI solutions on the backend and always keeping a human in the loop are over. Most retailers are comfortable bringing AI solutions directly to the customer.

“Keeping humans in the loop with AI is still sound advice, but the application has shifted,” Julie Geller, principal research director at Info-Tech Research Group, said in an email. “The question isn't about whether to keep humans involved; it is where in the customer journey that involvement matters most and how you design for it.”

Beyond revenue growth and improving the customer experience, e-commerce decisionmakers cited investing in AI for cost reduction and operational efficiency, according to the LogicBroker survey.

AI chatbots have a mixed impact on customer experience. While customers appreciate being able to access help 24/7 and fast responses; AI chatbots consistently fail to handle complex inquiries and customers often are frustrated by what they see as a roadblock to speaking with a human representative. 

“Many of these agents sit in an uncomfortable middle ground, neither convincingly human nor comfortably machine,” Geller said. “Responses can be hollow, nonsensical, or arrive too late to matter, and either way, the experience actively damages the brand it was meant to serve.”

With such experiences, it’s no surprise that two-thirds of consumers prefer talking to a human, according to a survey conducted by YouGov on behalf of Pegasystems Inc.

The problem is most companies are underinvesting in the handoff, Geller says. 

“When a customer reaches a live agent after an AI interaction, the burden of continuity should never fall on them,” Geller said. “They should not have to repeat themselves or wait while an agent catches up. The transition needs to be invisible, with the agent arriving fully informed and ready. Right now it rarely is.”

Meanwhile, the ROI of AI can be elusive. Gartner predicts that costs per resolution for generative AI will exceed $3 by 2030 — more expensive than many offshore agents. Rising data center costs, AI vendors pivoting from subsidized growth to seeking profitability, and complex use cases all may contribute to rising AI costs. 

That’s not to say companies shouldn’t invest in AI. Geller says the strongest use cases of AI is when it “is operating quietly in the background.”

Retailers are also investing in operational capabilities on the back-end: 44% are investing in pricing optimization, and 43% in automated inventory management, according to LogicBroker.

“Pricing optimization and demand forecasting work well because the variables are quantifiable, the feedback is immediate, and there is no direct customer relationship at stake if the model gets it wrong,” Geller said. 

AI can also be useful for product discovery, and consumers have signaled their comfort using the technology for research.

“On the customer-facing side, product discovery has real merit when trained on behavioral signals and context rather than just search terms,” Geller said. “A shopper who looks at something three times and never buys is communicating something, and AI reads that better than any keyword logic.”

Walmart hasn’t been shy about its AI ambitions. The retail giant recently announced an AI framework centered around four super agents, enabled purchases within ChatGPT and announced a new partnership with Google’s Gemini.

If it sounds like Walmart is moving fast, that’s intentional.

“For the last year or two, we’ve been tinkering with it,” Daniel Danker, executive vice president of AI acceleration, product and design at Walmart, said at the ICR Conference at the start of the year. “This is the year where tinkering becomes transformation. This is the year where we’ve built a level of mastery around that and we’ll start building things that deeply address customer problems.”

Danker has many examples of how AI could improve the Walmart shopping experience. For one thing, he sees repeat grocery purchases transitioning to a more automated model, with AI understanding how frequently to replenish items. For another, he envisions customers scrolling through AI-generated photos of themselves wearing the clothes they’re considering buying, or asking AI for recommendations based on what’s already in their closet. 

“I think of us all as carpenters, and we’ve all been using screwdrivers and suddenly someone shows up with a drill,” Danker said. “There are things you couldn’t build without power tools that you can with power tools. AI is a power tool for us.”

Danker also envisions shoppers doing a lot less scrolling across the board, thanks to the new level of personalization AI offers. That shift includes Walmart learning which products a customer wants to splurge on and which they prefer value offerings for; it includes narrowing a shopper’s search results to a specific product that might best fit their needs and suggesting others that complement their purchases.

“By the time you have the tomato paste and the ground beef and mozzarella, we're pretty sure you're making lasagna — and we don't need you to search eight times and scroll through many, many pages just to add the basil and the tomato sauce and the ricotta,” Danker said. “Those experiences are going to become more human, more connected. They'll understand your intent, and they'll serve it up to you much more easily.”

Walmart’s customer AI agent, Sparky, can already recognize if a specific shopper buys the same items every week and suggest a new order to that customer when they enter the app. It can field customer service questions and answer questions like, “What do I use to get this wine stain out of my carpet?” It can remind shoppers that they have an appointment at Walmart’s auto care center and also a prescription to pick up. 

Long-term, Danker believes that the Sparky chat interface and Walmart’s traditional search bar will collapse into one search system. But for now, the two exist separately.

“It's not going to happen overnight. It's also not going to happen all at once, but I do think this is a year where we will be delivering transformative experiences in commerce,” Danker said. “So you will definitely look back a year from now and say ‘Gosh, that's quite different than how I used to use Walmart. That's quite different than the expectations I used to bring when I opened an app.’”

What Danker is most excited about is the potential for Walmart to enter the chat as shoppers ask services like Chat GPT and Gemini questions that aren’t directly related to shopping. Customers who are unsure of their needs may ask these services about what type of TV would fit in their space, how to get that wine stain out or how to handle a new challenge that’s emerged with their newborn baby. Those, Danker argues, are questions that may not start with the intent to purchase, but could end up that way if Walmart is showing up in those spaces.

The big-box retailer’s integration with Gemini allows users to benefit from their membership perks with Walmart even if they order through the AI platform and it can also add items ordered with Gemini to a customer’s live Walmart shopping cart online, so they don’t have to place orders from multiple locations.

Of course, Walmart doesn’t have direct control over whether its product offerings show up in a Chat GPT or Gemini search, but Danker sees the big-box retailer as well-positioned for the algorithms. Specifically, Walmart’s focus on price and speed will give it a leg up if AI agents are making selections for shoppers, according to Danker.

Walmart’s stores will also see benefits from AI. Over time, the retailer plans to digitize the in-store experience and integrate some of those same AI-driven experiences into a physical environment. Associates are already using their own AI agent on the backend to assist with restocking shelves efficiently and prioritizing when things like spills happen. Walmart’s fulfillment centers are also using the technology to better predict what products they should store.

Not everything Walmart tries will be successful, but Danker is OK with that.

“The risk is that we build a few things that don't stick. I'd say there's a much bigger risk to not being out front,” Danker said. “We're going to lead because we think that it can do things for customers that we genuinely couldn't do before … the only way we'll get to the thing that works really well is by trying a lot of things along the way. And we're very lucky to be able to do that.”

In 2026, retailers are placing bigger bets on agentic AI commerce — even at the risk of losing direct access to customers and control over their data.

Recently Etsy, Target and Walmart have continued their move to bring merchandise offerings onto external platforms by partnering with Google’s Gemini and Microsoft’s Copilot.

All three companies joined forces with OpenAI's ChatGPT last year to make products available for purchase on the platform. That's on top of the efforts by Amazon and Walmart to popularize the AI consumer-facing assistants they built, named Rufus (which later became Alexa for Shopping) and Sparky respectively.

The rapid evolution of agentic AI — and the shift it poses to direct consumer access — is likely game changing for the retail industry.

“I kind of think that this is going to shake up retail just like the internet did,” Kartik Hosanagar, marketing professor at the Wharton School of the University of Pennsylvania, told Retail Dive.

Retail executives agree, as many took to the stage at the National Retail Federation’s 2026 Big Show in New York City to discuss the impact of AI.

Industry leaders like Ulta Beauty CEO Kecia Steelman and REI CEO Mary Beth Laughton told audiences at the event that AI will likely impact every aspect of the business.

The argument for joining forces with the likes of ChatGPT and Gemini is that retailers want to meet their customers where they are. Walmart's incoming CEO John Furner emphasized the sentiment while speaking at NRF.

AI-driven U.S. e-commerce traffic grew 758% year over year between Nov. 1 and Dec. 1, according to Adobe’s 2025 Holiday Shopping report. AI traffic to U.S. retail sites specifically on Cyber Monday increased 670%.

“What we expect is that consumers will deepen their engagement,” Katherine Black, a partner at Kearney leading food, drug and mass market retail, told Retail Dive via email. “More consumers will use it for shopping and for more types of shopping. As retailer capabilities become more prominent, that will encourage more usage as well.”

Meeting your customer where they’re at doesn’t necessarily come without a price, according to some industry experts.

Data ownership and the potential disintermediation of retailers are chief among the risks that external AI commerce poses to the industry in 2026.

Data and disintermediation

When a customer makes a purchase or casually peruses a retailer’s e-commerce site, a series of data points can be collected about their journey. Retailers can learn more about the needs and wants of a shopper by seeing what they searched for, which products they clicked to learn more about, how long they viewed each listing and more.

But what if that journey starts and ends within an external AI platform?

“It also shifts the locus of power,” Wharton’s Hosanagar told Retail Dive. “Whoever controls the agents now has the power.”

Deloitte’s 2026 Retail Industry Global Outlook found that 81% of surveyed retail executives think generative AI will weaken brand loyalty by 2027.

On stage at the NRF conference, Google and Alphabet CEO Sundar Pichai showcased a new suite of retailer commerce tools for its AI chatbot Gemini. 

The executive outlined how Gemini can take part in the customer journey, from discovery to decision and final purchase.

“He's talking about Google owning and collecting all of the data that happens around all three of those touch points,” Nikki Baird, vice president of strategy and product at retail technology provider Aptos, told Retail Dive. “And even if they share … if they're not sharing everything that happened that led to the discovery, that led to the decision, that led to the delivery, then the retailer loses so much context because it's outside of their realm.”

Pichai, however, insisted on stage that collaboration between Gemini and retailers was key.

“What we have learned from the last 27 years of working with retailers is that we only succeed together,” Pichai told the conference audience. “We want to use our full stack approach to help you shape this next chapter of retail.”

Risk goes beyond the potential loss of data about consumers, however. 

“The risk is that as these systems start to offer instant checkout and so on, I am doing all of my product research, discovery, understanding, including purchase not on Walmart's website, but on OpenAI's website,” Hosanagar said. “You're giving away the customer interactions, the brand experience, all of that, and then you become like a fulfillment company.”

This may be why Amazon has put so much effort into its own agent initiatives, Hosanagar added. 

The e-commerce giant debuted a dedicated website for its generative AI-powered Alexa+ assistant meant to help users research and plan out purchases conversationally. 

Still, participation in external AI platforms’ commerce initiatives may be a necessity and less of a choice moving forward. 

When OpenAI debuted its Instant Checkout shopping feature on ChatGPT in September, the company noted that the functionality could impact result rankings for queries. The tech company said it considers several factors — including the enablement of Instant Checkout, quality and price — when ranking merchants that sell the same product.

Looking at Walmart’s decisions, the company is hedging its bets by hopping onto external agentic commerce and aiming to become the retailer of choice on the new channel, according to Hosanager. 

“I don't see a retailer building their own agentic platform as a way to avoid participating in the agentic platforms,” Kearney’s Black added. “And of course, they would love for consumers to go to their brand directly and shop. But in reality, I think most retailers realize that there's a trend moving to these multi-retailer agents and so they have to be relevant in that space.”

Bringing product catalogs to AI chat platforms is just the beginning of the potential change to come in a customer’s shopping journey.

Early innings

About half of retail executives anticipate the collapse of the current multistep customer shopping journey by 2027 as the process shifts to a singular single AI-driven interaction, per Deloitte’s report.

The industry is just in the early stages of seeing e-commerce evolve with agentic AI, according to Hosanager.

“Where it gets interesting is where the buyer is actually using an agent to do the shopping, versus going to Gemini or going to ChatGPT and interacting with that system,” he said. 

For a frequent flyer who travels regularly for work, as an example, the future of agentic commerce may become more streamlined. An AI agent could learn the users preferences — such as preferred airlines and times of day — and proactively research the available flights that fit those needs. The AI agent could make the flight purchase on a customer’s behalf automatically, book applicable hotel accommodations and check the weather to help inform packing decisions — all without the user even opening the AI chat platform to get this process started. 

In the future, this scenario could play out between AI agents, where the user’s agent is interacting with an AI agent from the airline, for example.

A similar situation could play out in the world of retail, Hosanagar noted, and that could mean a drastic shift in how businesses interact with customers.

“For the retailer, you are increasingly going to interact with the end customer less and less, and you're interacting with their representative, which is the AI agent, which is now a new class of customer,” he said. “ It evaluates information differently. It needs information presented in a different format in order to understand it well. It can be persuaded, but it's persuaded differently than how you persuade humans.”

AI agents can also play a role in how customers shop in stores, according to Aptos’ Baird.

When Apple first launched the iPhone in 2007, retailers dealt with a massive change in how customers shopped in stores, Baird said. Consumers were searching online to check product reviews and competitors pricing while standing in store aisles.

Customers are now able to access the intelligence of ChatGPT right from their phones, which is providing more researched and instantaneous insights to them.

“It's not that you have the whole world wide web at your disposal while you're standing in the store,” Baird said. “It's that you have a store associate who's an expert in all stores.”

The moment may call for equipping employees with their own AI technology in stores so they can get instant insights about a customer they’re interacting with, such as their clothing preferences or whether or not they’re a frequent product returner. 

Or a retailer’s own AI agent could proactively reach out to customers on behalf of a store associate to inform the consumer when a product they like is back in stock at the location, encouraging a sale.

“The point is to let the store associates shine,” Baird added.