How generative AI is transforming online retail content and personalisation

AI in e-commerce isn’t breaking news. It’s been around for a while—just not the shiny, chatty version we’re seeing now. Back in the early 2010s, deep learning quietly slipped into conversion rate optimisation tools like Dynamic Yield and Optimizely, tweaking landing pages and layouts behind the scenes.

But with the recent arrival of powerful Large Language Models (LLMs) and Diffusion Models (the magic behind those eerily perfect AI-generated images), things escalated quickly. Suddenly, e-commerce had access to a toolbox that could generate, translate, tweak, and visualise—all at once. What used to take teams of people and months of production can now happen in a few clicks. The result? A gold rush of AI-powered startups, all vying for a place in your tech stack—and your product pages.

Less busywork, more business

Launching a product online used to be a logistical ballet. You needed high-quality photos, well-written descriptions (preferably in five languages), SEO-optimised copy, and campaign content to match. Coordinating this across teams and timelines often meant delays, missed windows, and a lot of stress.

Now? AI is stepping in—not to replace people, but to relieve them of the repetitive bits. Tools like Pixelcut and Flair.ai let you drop a product into any background you want. Fashion brands without access to models or studios can now dress up AI-generated humans in their latest looks (shoutout to Botika and Lalaland).

Text-wise, LLMs like ChatGPT and Claude can spin out a product description in moments—and with the right prompt, even help with tone, SEO, and localisation. Platforms like Hypotenuse.ai bundle these capabilities into one neat workspace. And for multilingual brands, AI-powered localisation tools (like Lokalise.com) offer cleaner translations, fewer spreadsheets, and much less coffee-fuelled late-night editing.

Yes, there’s a catch (of course there is)

Here’s where it gets nuanced. These tools are powerful, yes—but not perfect. Hallucinations (when AI makes things up) can slip into copy. Scammy sellers can flood the web with fake products using AI-generated imagery, eroding consumer trust. And let’s not forget: by default, AI tends to average things out. That means safe, generic content—not exactly the stuff strong brands are built on.

The good news? AI follows orders really well—especially good ones. That’s where human expertise comes in. Strategy, taste, brand voice, aesthetic judgement—these aren’t things you can outsource to a machine. At least, not yet.

We’ve found that AI works best as a force multiplier for creative professionals. It accelerates the process, but you still need someone steering the ship. Someone with criteria. Someone who knows when a headline sings or when a photo just clicks.

From A/B to AI/B

Personalisation isn’t a new idea in ecommerce. For years, brands have used A/B testing and behavioural data to create segmented experiences. But these were still limited by what you could physically produce—alternate headlines, maybe a different photo here or there.

Now, generative AI makes it possible to create variations at scale. And this changes everything.

Imagine a visitor to your fashion site always clicks on models that resemble a certain look—say, dark hair, sporty vibe, casual setting. The site could adapt, showing AI-generated product photos that match that customer’s preferences. Not only is this feasible, it’s already starting to happen.

And it’s not just about faces. Context matters too. When we helped redesign the Victorinox website, we saw how perceptions of the Swiss Army Knife varied across markets. What’s practical in one region may be aspirational in another. AI-generated imagery can help tailor those visuals to cultural cues—no extra shoots or budgets required.

Of course, this kind of hyper-personalisation still needs guardrails. Someone has to decide what’s appropriate, what’s on-brand, and what makes sense for the customer. In other words, humans still hold the compass.

Crafted by humans, assisted by machines

In the rush to embrace AI, it’s tempting to go all-in. But at Elespacio, we believe in balance. Yes, we use AI to speed up processes, scale content, and unlock new possibilities. But we don’t hand over the keys.

We blend AI’s potential with human intuition to craft e-commerce experiences that resonate. The result? Brands that speak clearly, move quickly, and stand out—without shouting.

Because while AI is smart, it doesn’t know your audience the way you do. It doesn’t understand nuance, timing, or tone. It can generate, but it can’t feel.

That’s where we come in.