Back in 1997, the term ‘Search Engine Optimisation’ (SEO) first became commonplace in the lexicon of the internet. Along with the advent of search engines themselves, those three little letters fundamentally reshaped digital browsing habits, and with them the very history of online retail.
For thirty years, SEO has formed the building blocks of digital discoverability. If you want, say, a purple hoodie you’re selling to actually be discoverable for consumers who are in the market for a purple hoodie, SEO has been the undisputed solution to connect shopper and shop. The vital thread between intent, solution and actionable outcome. As such, retailers have made it their mission to get the process of search engine optimisation down to a fine art.
Like most algorithmically-fuelled endeavours, SEO has quantifiable measures of success. If you literally ask Google what the fundamentals of SEO look like, it’ll tell you that “key pillars include comprehensive keyword research, on-page optimisation (titles, meta descriptions), creating valuable, user-focused content, ensuring fast, mobile-friendly technical performance, and building backlinks”.
So, if you’re wondering what all the fuss around GEO (Generative Engine Optimisation) is about, it’s for the simple reason that these rules for discoverability, which have been unquestionable tenets of success in ecommerce for three decades, are starting to falter. And the culprit is AI.
Before retailers start to strip out their SEO investments and throw them to the wolves, it’s worth noting that the results aren’t as dramatic as some sources might claim. In fact, currently, the chunk of traditional search’s pie that AI is biting off is sitting at under one per cent. However, its usage as an entry point to ecommerce retailers has grown dramatically in a short space of time. In the UK, it’s increased from 0.44 per cent in Dec 2024 to 0.89 per cent in Dec 2025, according to research from Datos.
When you consider that AI became a feasible alternative to traditional search in just 2022 with the launch of ChatGPT, and that as of 2026, this one solution alone has over 800 million weekly users, it’s clear to see why that 0.89 per cent change is causing some serious discussion.
The argument for many will be that, just as an entry into search optimisation has proven to be a sound investment, as AI continues to grow at pace, and with the much-touted agentic shopping explosion imminently expected, preparing for a world where SEO is no longer the only ‘engine optimisation’ on the agenda is very important.
What makes this moment particularly interesting, though, is that GEO is not emerging as a clean replacement for SEO in the way some commentators would like to suggest. Instead, it layers new behaviours on top of old ones. Consumers haven’t stopped using search. What they’ve done is added another route into process of discovery. The difference is that this new route behaves less like a directory of potential solutions and more like a recommendation, weighting the sway and discoverability of GEO considerably over its SEO counterpart.
Traditional search presents options and leaves the consumer to do the work. Generative AI compresses that process. It interprets intent, filters choices, and presents an answer that’s designed to feel ‘considered’, rather than ranked. In leu of a quick search for ‘purple hoodie’, a prospective shopper may now turn to AI and say: ‘I’m 5 foot 9, I’m 11 stone and I want a purple hoodie that will look good on me’. Discovery now shifts from browsing to the delegation of browsing to a smart robot.
For retailers, visibility is therefore no longer just about appearing high on a list of links, but about being included in the reasoning behind an answer. In practical terms, that means the content that once existed to drive clicks increasingly exists to inform AI systems themselves. The outcome may still be a visit to a retailer’s website, but the moment of consideration has already happened elsewhere.
Chris Holyland, digital and omnichannel director at Currys, tells Retail Gazette how quickly that reality became apparent. A year ago, he says, much of the industry’s concern centred on disintermediation; the idea that tools like ChatGPT would remove retailers from the journey entirely.
Instead, the opposite has happened. Customers are conducting research through AI platforms, but still being directed back to retailers to complete the purchase. The implication is that retailers need to ensure they’re visible at that earlier stage of decision-making. As Holyland puts it, the focus has become making sure “the LLM recognises us and also recommends us”.
That shift sounds straightforward, but it introduces a tension that retailers are only just beginning to grapple with. For decades, success in digital commerce has been measured through traffic. More clicks meant more opportunity. GEO introduces a scenario where content may influence a purchase without ever generating a visit.
Information that once lived on a retailer’s site is now summarised and surfaced directly within AI overviews or conversational responses. The value of that content remains, but its attribution becomes less obvious.
In other words, the industry is edging towards a zero-click reality, at least in parts of the journey. And that requires a recalibration of what success looks like. Being the source of the answer may matter more than owning the interaction itself.
This is partly why the conversation around GEO often feels confused. It’s being framed as a marketing problem, when in reality it’s easily as much an operational one. Generative systems reward clarity, consistency and structured information. They surface brands that explain themselves well, whose product data is clean, whose information is easy to interpret, and whose presence across the wider web reinforces credibility.
Melissa Minkow, global director of retail strategy and insights at CI&T, argues that this is where many retailers still face a fundamental challenge. “Data is the foundation for AI,” she says, noting that many organisations are still working with fragmented or poorly structured information. The risk is not simply missing out on AI visibility, but missing out on the broader efficiencies that come with having data in a usable state at all. As such, wielding agentic AI as a truly effective tool requires fixing long-standing structural issues.
There’s also a behavioural element at play. Consumers are becoming increasingly mission-oriented, seeking faster resolution rather than longer engagement. For years, time spent on site was treated as a proxy for success. Increasingly, the opposite may be true. The better the decision support, the shorter the journey needs to be. AI simply accelerates that expectation.
None of this suggests that SEO suddenly loses relevance. The fundamentals that have underpinned search visibility for decades, such as authority, relevance, technical performance, useful content, all remain intact. If anything, they become more important. Large language models draw from the same ecosystem of information that search engines have long relied upon. Poor content doesn’t magically become more effective simply because it’s processed by AI.
What really changes is the emphasis. Keyword optimisation is less vital than context. Product descriptions evolve from functional specifications towards use cases and intent. Content needs to answer questions that are actually asked, rather than those that neatly align with marketing taxonomy.
Looking slightly further ahead, the conversation inevitably turns to agentic commerce, the idea that AI systems won’t just recommend products, but act on behalf of consumers. Here, the impact is likely to be uneven. High-consideration purchases will remain human decisions for some time yet. But replenishment categories, commoditised purchases and repeat buying behaviour are far more susceptible to automation. The moment an agent can reliably source the right product at the right price with minimal friction, brand visibility at the recommendation stage will be critical.
Which brings the discussion back to realism. GEO isn’t a silver bullet, nor is it an overnight transformation. The share of traffic it represents today remains small. But its trajectory mirrors previous shifts in digital behaviour. Slow at first, then suddenly unavoidable. Retailers that treat it as speculative risk missing the point entirely. Embracing GEO in early 2026 is acknowledging where consumer attention is gradually moving.
SEO connected intent to outcome for three decades because it aligned with how people searched. GEO is building momentum for the same reason. As consumers grow more comfortable asking questions instead of typing queries, retailers need to ensure that they’re easy to find, easy to trust, and easy for both humans and machines to understand.
Click here to sign up to Retail Gazette‘s free daily email newsletter
