First off, we’ll apologise in advance. Yes, this is about AI.
We know. Everyone’s talking about it. Every tool has it. Every sales deck leans on it. At this point, it’s less of a trend and more of a background noise you can’t switch off.
But that’s exactly the problem.
Somewhere between the hype, the headlines, and the “AI-powered” labels slapped on everything, the actual value has got a bit lost, especially in eCommerce.
So it’s worth asking a simpler question: do you actually need AI, or do you just need to get the fundamentals right?
AI isn’t magic, it’s just pattern recognition
Strip away the buzzwords, and AI in eCommerce is pretty straightforward: it looks at data, finds patterns, makes predictions & automates actions at scale.
That’s it.
Useful? Yes. Transformational? Not on its own.
AI is entirely dependent on what you feed it, so if you’ve got bad product data, messy category structures, or poor UX, AI won’t magically improve it just because an AI tool is in place.
The biggest lie: ‘AI will grow your store’.
AI is being sold as a growth lever. A shortcut. Something you can switch on and instantly scale revenue with. You can’t.
AI doesn’t fix underlying problems. It amplifies what’s already there.
If your store has a clean structure, well-tagged products, strong navigation, and a smooth checkout journey, AI can enhance and scale that. But if your store has inconsistent product data, poor categorisation, or weak UX, AI will simply make the chaos happen faster.
Many brands jump into AI expecting a shortcut, only to see zero measurable impact. That’s because the foundation - the fundamentals - were never in place.
SEO, AI & Discovery: Nothing has really changed
You can see the same pattern in how AI is influencing search & discovery.
With AI overviews and chatbots answering queries, it can feel like you’re missing out if you’re not stressed about it.
Thing is, though, nothing has changed.
These systems still rely on the same fundamentals as traditional search: crawlable site structure, clear content hierarchy, strong internal linking & trust signals.
For eCommerce, this means:
- Properly structured categories that make sense to both humans and algorithms
- Product pages with clear descriptions & useful content
- Clean product data and schema markup
- Logical navigation that reflects real shopping journeys
Paul Gray, our SEO lead, said
"AI is a powerful, useful tool for SEO, from research and blog briefs to structuring product content and shaping meta titles and descriptions. But it shouldn’t replace the thinking. Use it to speed things up, not to do the job for you."
Basically, if you’ve done your SEO properly, AI should already be picking up your site. Visibility in AI-driven search is a byproduct of doing the basics well.
The buzzwords
Most AI features sound more advanced than they actually are…
AI Personalisation: Usually, segmented experiences with some automation layered in.
Smart Recommendations: Driven by historical behaviour. Only as good as your product tagging.
AI Search: Improved matching, heavily dependent on clean data.
Predictive Analytics: Forecasting based on past performance. Not a crystal ball.
AI Content Generation: Scales output. Doesn’t guarantee quality or differentiation.
None of this is useless, but none of it is a replacement for strategy.
Where AI works best in eCommerce
Used properly, AI can drive real impact. But only in the right places:
- On-site search & merchandising
- Email & SMS optimisation (especially in tools like Klaviyo)
- Product data enrichment
- Customer support assistance
- Faster reporting & analysis
AI works best where tasks are repeatable, high volume & structured.
Where it falls apart
Most eCommerce brands aren’t ready for AI - not because the tools don’t exist, but because the foundations are weak. Product data is inconsistent or incomplete, categories don’t follow a logical structure, and platforms and systems are often disconnected. Customer segmentation is patchy, tracking is unreliable, and reporting doesn’t reflect reality.
When AI is layered on top of these issues, it doesn’t fix anything. It simply amplifies the problems. A poorly structured search algorithm won’t suddenly deliver perfect product discovery; it will just show the wrong products faster. A recommendation engine fed with incomplete or incorrectly tagged products won’t increase basket size; it will confuse customers. AI, in other words, exposes cracks rather than covering them up.
For example, a brand installs AI-powered search on top of poorly tagged products. A search for “black dress” returns hundreds of results, many of which are irrelevant. Conversion doesn’t improve. In some cases, it drops.
The issue isn’t the tool. It’s the data feeding it.
The modern stack problem
Adding AI to a messy eCommerce stack is another common pitfall. Many brands now use a platform like Shopify or Magento, a search tool, a personalisation layer, an email platform, and an analytics suite - each claiming to use AI in some way.
The result is often overlapping functionality, conflicting data, higher costs, and slower teams. More tools do not equal better performance. Clean integration, strong data, and a clear strategy do.
How to use AI without wasting budget
The biggest and easiest tip we can give you is keep a human in charge. Merchandising, brand voice, and strategy can’t be outsourced to an algorithm. AI should augment these functions, not replace them.
A simple test before implementing any AI tool is this: ask what specific problem it solves, what data it relies on, whether that data is reliable, how it fits into the existing stack, and whether its impact can be tied to revenue. If the answer is unclear, it’s probably just an expensive test.
The takeaway
Get your fundamentals right first.
When your SEO, product data, and systems are solid, AI becomes a powerful tool for growth. If they’re not, AI won’t save you - it’ll just make the chaos more visible.
P.S. Want to download our full guide on SEO for AI? Check it out here.
