February 23, 2026
If you spend any time reading industry updates like we do, then there's no doubt you’ll be familiar with headlines declaring the 'death of search'. Generative AI is getting better, the sky is falling, and traditional SEO is supposedly obsolete. However, when you look at the actual data, it tells a different story. The landscape of commerce is shifting, there’s no denying that, but it’s not what recent posts suggest.
So, to give you a more accurate overview of AI and search, we’re taking a look at some of the most recent stories and events, and what they actually mean for your digital strategy.
If people are still searching, how does AI actually fit into things and how is it different? An insightful interview with Perplexity AI recently broke down the crucial difference between traditional search engines and true AI search.
Classic search indexes whole web pages (called whole-document processing), whereas AI search relies on sub-document processing. Instead of pulling fifty full articles to rank, the engine breaks text down into smaller chunks and retrieves tens of thousands of specific snippets to feed the AI's context window. This gives the model a huge base of facts to work with, and the idea is that this should reduce hallucination.
For marketers, this is about the shift toward what some are calling ‘Answer Engine Optimisation’ (AEO). The goal is no longer just ranking a landing page, but structuring your information (pricing, descriptions etc.) clearly so the AI pulls your snippets out when giving the definitive answer.
There is a wild claim circulating that AI has slashed organic search traffic by up to 50%. But a study from Graphite looked at quantitative panel data from thousands of the largest US websites and found something entirely different: organic search traffic is only down by about 2.5% year-over-year. So there’s really nothing worth worrying about.
But why the huge discrepancy?
Well, most doom-and-gloom numbers you may have come across are mostly based on self-reported surveys, which are notoriously bad at measuring actual digital behaviour. The reality is that traffic to search engines remains pretty stable. In fact, 90% of clicks from Google still go to organic results rather than ads. People are still searching, and traditional SEO is still driving the largest share of reliable traffic.
As AI platforms secure their massive user bases, they are of course turning to monetisation, because the power required to run these LLMs isn’t cheap. OpenAI is officially rolling out ads in ChatGPT, but they aren't playing by the usual performance marketing rules.
Early reports indicate OpenAI is asking advertisers for $1 million trial commitments and $60 CPMs (cost per thousand impressions), charging for views rather than clicks. Without any kind of self-serve platform yet (like the Google Ads interface), this is just high-end brand awareness advertising. It's almost like paying for a television advert. And the ads aren’t interwoven into the response either, they sit separately below. This raises the question - if a user is focused on reading a response, how much attention will they pay to the actual ad? It is an expensive experiment and reinforces the idea that traditional paid search remains the more reliable channel.
We talk a lot about AI helping customers find products, but what happens when the AI tries to buy them? ‘Agentic shopping’ is supposedly the next big thing, however eBay recently updated its user agreement to explicitly ban AI bots from executing purchases. Bots have always been banned, but LLMs are now specifically part of that.
This is an important signal for eCommerce. When a bot buys a product autonomously, the chain of attribution is blurred. It also introduces legal headaches regarding refunds, fraud, and user consent. AI is highly effective for the research and discovery phase, but eBay’s move suggests some major platforms are drawing a hard line - the final checkout decision requires human intent.
While eBay is restricting autonomous buying, Shopify is busy trying to make it easier. Shopify recently shared more details on their Universal Commerce Protocol (UCP), a framework co-developed with Google that lets merchants decide how AI agents see and interact with their website and products. The idea is to reduce friction and make it easier for agents to do what they’re being asked.
Shopify’s leadership calls this a move toward ‘merit-based shopping.’ In this model, an AI recommends a product because it perfectly matches a user’s specific context, not just because the brand has the biggest advertising budget. This supposedly levels the playing field, allowing highly relevant challenger brands to compete with established giants.
Most importantly, Shopify is opening this up so even non-Shopify merchants can syndicate their catalogs to ensure their products are discoverable by AI agents.
The bottom line is that organic search is, and continues to be, the backbone of digital visibility, while AI is carving out new and more specific channels for discovery. The brands that will earn success over the next couple of years will be the ones that maintain strong technical SEO while formatting their data clearly enough to be retrieved by AI agents.
If you’re looking for expert help navigating this changing landscape, don’t hesitate to get in touch with us today.
