


If there was one thread running through Shoptalk this year, it was how quickly AI is moving from concept to reality in retail.
Not in a distant, experimental way, but in how consumers are already discovering products and navigating the shopper journey.
The shift is more than just new tools. The fundamentals of retail are also starting to change: where discovery happens, how decisions get made and who stays closest to the customer.
Across sessions and tech labs, from leaders at Lowe’s to Anthropologie, the same patterns kept coming up at Shoptalk.
Shopping isn’t only starting on a retailer’s site or even a marketplace. Increasingly, parts of the journey are happening earlier and elsewhere like inside AI tools, answer engines and conversational interfaces. In some cases, the browsing phase is getting shorter. In others, it’s being skipped entirely.
The shift is already starting to play out. For example, Sephora announced at Shoptalk a partnership with OpenAI to bring its experience into ChatGPT, an early sign that discovery and decision-making are starting to happen directly in AI-driven environments.
This changes how demand is captured. If fewer shoppers begin their journey on owned properties, then traditional entry points—homepage visits, category browsing, search—become less reliable as the starting point for engagement. The implication is subtle but important: retailers and brands need to think more proactively about how they can create more demand.
AI is starting to play a more active role in helping people choose what to buy, going beyond surfacing options and into decision-making.
That puts more weight on relevance and trust. If the experience feels helpful, people lean in. If it doesn’t, they move on quickly.
As Lowe’s Joe Cano, SVP of Digital Commerce, shared on stage at Shoptalk, that’s already visible through Mylow, Lowe’s AI shopping agent, which helps customers navigate everything from project questions to product selection in real-time, bringing discovery and purchase into a single experience.
As discovery and decision-making move into new environments, control starts to shift with it.
Retailers and brands risk losing:
If retailers and brands are not present in the moments where decisions are being shaped, you show up later when fewer decisions are left to influence.
Put another way, the closer you are to the start of the journey, the more opportunity you have to shape outcomes. The further downstream you are, the more you’re competing on price and availability.
This is where the tone of the conversations at Shoptalk got more real.
There’s a lot of excitement around AI, but also a growing recognition that this shift comes with tradeoffs.
At the same time, expectations are rising. Customers want experiences that feel tailored, helpful and intuitive. They’re more willing to engage when it works and quicker to disengage when it doesn’t.
So the challenge becomes clear: how do you stay visible, relevant and influential as the journey moves beyond your walls? And more importantly, how do you do that in a way that still feels cohesive, rather than fragmented across channels and touchpoints.
One of the most consistent themes across sessions was the need to think beyond isolated channels.
It’s less about individual campaigns or placements, and more about how the whole journey fits together. Retail media sits right in the middle of this shift.
Not just as a way to monetize traffic, but as a way to stay present as that traffic fragments.
If discovery is happening before someone reaches your site, then reaching them earlier matters more.
That’s where offsite media becomes critical—connecting with potential customers before AI or other platforms narrow down their options.
It also allows retailers to reintroduce themselves into moments they don’t directly control, helping maintain visibility even as the starting point of the journey shifts elsewhere.
It’s not enough to reach people. Retailers and their supplier brands have to be relevant when it matters.
That comes down to how well data and decisioning are working together, so messaging reflects what someone actually needs in that moment.
That point came up at Shoptalk during a technical lab with Yum! Brands’ Gyan Mathias, who talked about the need to make data “machine-readable for AI”—a reminder that these experiences don’t work without the right foundation in place.
This is where a lot of the conversation around personalization is starting to evolve from static targeting to more dynamic, responsive engagement that adapts based on behavior and contextual data.
If the journey spans multiple environments, your presence has to as well.
Retailers are being pushed to operate more like media platforms where they can show up across channels beyond their own ecosystem. This shift requires not just reach but coordination. Without it, the risk is duplicative messaging, wasted impressions and a disjointed experience for the customer.
A lot of the conversation at Shoptalk focused on the gap between what AI can do and what actually works in practice.
The difference comes down to execution. It’s one thing to build models. It’s another to make them actionable and actually connect them to real data, real decisions and real outcomes.
This means integrating AI into the systems and workflows that actually drive marketing and commerce rather than treating it as a standalone capability.
As Ryan Den Rooijen from Currys pointed out onstage, you can’t just plug in AI and expect it to work. “Models can reason… but they won’t do anything—they don’t have the muscles to actually act.” You need the right data, the right context and the ability to activate it and create business value.
What stood out most over the course of Shoptalk wasn’t one specific trend. It was a broader shift in how retail operates:
This creates real pressure but also a real opportunity. The retailers and brands that adapt will be the ones that stay connected to the customer, even as the path to purchase becomes less direct.
This is where a lot of the conversations we’re having with clients are focused right now:
That’s where your entire media mix—from programmatic to retail media—becomes more than a tactic. It becomes part of how you navigate the shift by staying present in the moments that shape the shopper journey.
As these changes take hold, one thing becomes clear: success depends on how well everything connects. Data, identity, media and decisioning can’t operate in silos if the goal is to stay relevant across a more fragmented journey.
That’s where Epsilon Retail Media is spending a lot of time with retailers and brands right now.
Helping them:
The retailers and brands that stay ahead will have anchored AI in a strong identity foundation that makes every decision smarter and every interaction relevant. To dig into best practices concerning identity-driven AI, take a look at the recent IDC White Paper, Improve data quality to support quality AI outcomes, sponsored by Epsilon.