AI is getting louder in ecommerce.
So is GEO. So is TikTok Shop. So is the promise that agents will soon buy on behalf of customers.
But in Damian and Matt’s conversation, the real signal is quieter.
Customer behavior still decides what wins.
This episode moves through European ecommerce growth, the decline of offline shopping, early GEO lessons, the limits of native AI checkout, the rise of social commerce, and what Uxtivity is building in response to growing “marketing agency fatigue.”
Damian opens with what sounds like a small change.
“78 % of European Internet users bought online in 2025.”
Compared with 2024, it is “only 1 % of difference.” But that is exactly why the number matters. The percentage looks calm. The behavior underneath it does not. Damian immediately translates the stat into something operators can feel: “it’s still over 5 million people more” buying online.
That shifts the conversation away from headline inflation and toward business reality.
More people are entering or deepening the habit of buying online. And when Damian looks at where that movement is strongest, he does not treat Europe as one market. He points to Romania, Italy, and Bulgaria as the three directions worth watching most closely if a brand is thinking about expansion.
That framing matters.
Because growth alone is not enough. Damian adds a useful qualifier: not all ecommerce growth flows to branded stores. Sometimes the lift goes to marketplaces instead. He mentions Amazon, Shein, and Temu as reminders that growing online demand does not automatically mean growing direct-to-consumer demand.
One of the sharpest parts of the conversation is not a statistic at all. It is Matt’s lived explanation of why ecommerce keeps getting stronger.
“Every single year, the offline experience is getting worse and worse and worse,” while “the ecommerce experience… is getting better and better and better.”
That is not theory. It is a diagnosis.
Matt describes empty stores, vacant locations, and a retail experience that often feels abandoned even when the doors are still open. The problem is not only convenience. It is care. In his words, “the customer experience is gone. No one really cares about you anymore.”
That gives a very practical explanation for why more people keep moving online.
Not because ecommerce is shiny.
Because for many categories, it increasingly feels like the more usable option.
Damian adds another layer here through the lens of social and live commerce. Offline and online do not always have to fight as separate worlds. They can also blend. He points to models where physical inventory and live shopping behavior come together, especially in the US, where that behavior seems more mature than in Europe.
The GEO section is strong because it avoids two bad extremes.
It does not dismiss GEO.
And it does not worship it.
Damian’s position is clear: “GEO, Generative Engine Optimizations, tactics and strategies… are still very early in the process.” That means experimentation makes sense. Blind overreaction does not.
His next point is even more useful for marketers.
“SEO is still very valid.” GEO may take “a little credit” from SEO, but it does not replace the need for strong fundamentals.
Then comes the distinction that most people miss when they talk about “being visible in AI.”
“Being visible in chat GPT is not the same thing as getting cited.”
That one sentence changes the whole strategy.
If citation behavior differs from general visibility, then brands need to think harder about structure, clarity, and what information appears early on the page. Matt picks up on this through the phrase “atomic facts,” calling it “a really cool term to use in copywriting.”
From there, the conversation moves into a question that matters well beyond GEO: how do you answer the most important question in the first third of the page without hurting UX?
Damian’s answer is simple and sharp. Most companies make the top of the page about themselves. They write what they know best. But that is not what the customer needs first. The fix is to structure the copy around “the customer point of view” and address “his context, his problem, his desires, his doubts.”
That is good copywriting. And increasingly, it is good discoverability too.
When the conversation turns to OpenAI’s checkout direction, Damian puts the emotional subtitle on it immediately:
“We told you so.”
His reasoning is not anti-AI. It is behavioral.
He explains that when customers are researching products, they are often still comparing stores, products, and prices. They are not yet in a buying state. So the idea of collapsing that process into native checkout inside ChatGPT never matched the underlying journey as neatly as the hype suggested.
He says it plainly:
“When customers research or look for products, they are still not ready to buy the product.”
That is why his own experiments with AI product discovery have felt limited. He says he has tried looking for several products in ChatGPT and got “pretty lousy results.” He is especially skeptical when the product choice depends on pricing accuracy and nuanced comparison.
The bigger takeaway is stronger than the product critique.
“The most important part in any business and especially ecommerce business is to follow your customer behavior.” And the behavior signal, in his view, is clear: “people do not want or do not need even to buy from chat GPT directly.”
Matt agrees with the broad direction, but adds a useful refinement. AI can still be very valuable in commerce, just not always where the loudest people want it to be. He sees it as a research layer. Something that helps evaluate ingredients, materials, and combinations before a decision is made.
That is a much more believable role. Not the cashier. The assistant before the cashier.
The next thread pushes the same idea further.
What about agents buying for customers?
Matt captures the feeling well. The speed at which people have normalized AI is surprising. The idea that a person might give one prompt and then wait for the product to show up at the door is, in his words, “mind blowing.”
Damian does not dismiss the possibility. He uses AI heavily himself.
“AI is my ally in my business. I really do use it in many tasks.”
But he does challenge the timing and scale of the promise.
“At this point, AI agents shopping for customers will be very niche.”
Why?
Because trust, need, and real capability do not move at the same speed as tech headlines.
He uses a very practical example. When researching printers, ChatGPT suggested one that was “$50 more expensive than in other store.” Then he turns that into a simple truth that many shoppers would share: “I’m not fond of losing $50 in sake of saving five minutes of my shopping experience.”
That line matters because it grounds the whole conversation. People do not adopt buying behavior because the workflow sounds futuristic. They adopt it when it feels safe, useful, and better than the alternative.
For now, AI is much closer to being a strong research companion than a trusted autonomous buyer.
The TikTok Shop section matters because it shifts the lens from interface to infrastructure.
Matt reacts to the warehouse story with genuine surprise, then quickly lands on the strategic point: it looks like “a pretty smart play.” He cites “$23 billion in volume in sold products” as part of why the move deserves attention.
But Damian does not reduce the story to size alone.
What interests him is the behavioral model underneath it. He says marketplace shopping behavior and social-media shopping behavior are not the same. “These are two separate kind of behaviors.”
That is a useful lens.
If the behaviors differ, then TikTok Shop is not simply trying to copy Amazon. It may be building around a different mode of intent, discovery, and purchase.
At the same time, Damian keeps his feet on the ground. He wants to see how far TikTok can really go on the operations side, because Amazon’s fulfillment edge was built over time. “This is something Amazon was and is building for the last 20 years, right?”
So yes, it is a serious move.
But there is still a long distance between entering the game and owning the infrastructure game.
The sponsor segment closes the episode, but it does not feel disconnected from the rest of the discussion.
If the whole conversation is really about trust, behavior, and tangible outcomes, then the final announcement fits perfectly.
Damian introduces a new performance-based retainer model for Uxtivity’s ecommerce growth program. It can be “mostly based on a success fee aligned with individually predetermined revenue levels” for the ecommerce channel.
The reason is brutally familiar.
“Marketing agency fatigue” is real.
Brands are tired of paying large fees and not seeing real business impact. Damian says that directly. The point of the new model is to answer that trust problem with a structure that feels more like partnership and less like overhead.
If there is one idea that ties the whole episode together, it is this:
AI does not remove the need for ecommerce fundamentals. It makes weak fundamentals easier to expose.
The brands that win still need clear positioning, strong product communication, useful page structure, and decision-making support that matches the customer’s stage.
That is why Damian keeps returning to behavior. And why Matt keeps returning to experience. The toolset is changing fast. The human logic underneath it is changing much more slowly.
Subscribe for more episodes of Coffee with Ecommerce, your weekly shot of hot trends and practical tips from the ecommerce world.
