AI and the evolution of search behaviour

Jay Tanel at creative agency Marks describes how brands must adapt to changing search behaviours

 

Most people are only talking about the tip of the iceberg when it comes to AI’s impact on search engine optimisation. Your LinkedIn is probably full of it: “SEO is dead”, “SEO is evolving”, the impact of Amazon Rufus, Agentic Shopping, AEO (Answer Engine Optimisation) and GEO (Generative Engine Optimisation) and so on.

 

So what lies beneath the surface of these buzz terms that are crowding our feeds?

 

 

Short-term potential chaos, long-term gain

The reality is that AI is going to mess things up in the world of online search. Likely very publicly. For example, some businesses will wake up one morning to find an AI output live on page one of their search results claiming something absolutely wild. It’ll happen.

 

A recent HBR article pointed out that AI “tends to make things up, leave things out and create so many possibilities that it’s hard to figure out which ones will be effective”. But remember, humans have been making search engine and listing mistakes for years: keyword stuffing, deliberately using inaccurate terms, growth-hacking gone wild.

 

So, my take is this: AI absolutely makes mistakes faster, but it also scales learning and correction faster than humans ever could.

 

In the long term, it will be better for businesses, better for brands, better for consumers. Here are the main reasons why…

 

 

From keywords to context

Traditional SEO (TSEO) was simple: shoppers typed two to three words into a search bar, and if you indexed and ranked well, you’re in the green.

 

But AI has changed the language of search. Prompts are averaging 10 to 11 words today. That means context now matters as much as the specific keywords. On top of that, it’s not a single search and move on; we’re talking about a conversation, with at least one follow-up prompt or clarifying question. That flips the volume ratio: context now will soon outweigh keywords.

 

Your choice of language (not just keywords) becomes AI + SEO fuel. Your brand voice, tone, aspirational situations, and brand signals can be used to lead shoppers to convert on your listing.

 

And external context doesn’t just come from starting prompts to kick us off in the shopper journey. It’s baked into consumer post-purchase reviews and middle-funnel Q&As. When someone writes “love the flavour, hate the packaging”, that’s valuable context that an AI model will remember and eventually rank you on.

 

If your search team is still treating keywords like a magic bullet, you’re already behind.

 

 

Amazon has an AI team, not a single player

We need to remember that AI-powered search is a game of many players. For example, while Amazon Rufus might feel like everyone’s favourite golden retriever of AI, it’s not the only one making moves.

 

Rufus is a shopper-facing agentic AI that remembers search history, purchases, and preferences. It can basically recommend products based on shoppers’ previous behaviour.

 

But you also have Amazon’s Catalog AI, the structural backbone of the retailer’s product data, as well as COSMO, Amazon’s operational AI powering behind the scenes in comprehension. Then there’s Alexa+ (close cousin to Rufus), which is a shopper-facing context machine turning spoken queries into context-rich data.

 

Other players include Walmart’s Spark, the newer competitor bringing its own flavour of agentic shopping; and of course ChatGPT, the AI ‘got-to’ for shoppers today, according to many (more than half of Brits now use ChatGPT to help with their shopping, according to new research from Omnisend).

 

So these aren’t just buzzwords. They’re distinct players, each with unique strengths, weaknesses, and roles. If AI + SEO used to be a simple game of catch, it’s now a coordinated team scrimmage.

 

 

We have a new metrics problem

The catch with these developments is that our dashboards haven’t even come close to catching up with the complexity that AI + SEO offers. Clicks, bounce rates, keyword rankings: those are fine for traditional search engine optimisation. But AI is indexing, interpreting, and referencing differently.

 

This means we need new KPIs that should look something like this: 

• Context quality: Is your content around current keywords clear, consistent and accurate?
• Sentiment signals: Does shopper feedback compete with brand identity? (If so, AI will notice)
• AI versus human traffic: The distractable human shopper is now measured alongside the multimodal AI system that indexes everything and doesn’t get bored.
• Share of voice: In addition to keywords, we need to consider the context around them. 

If you’re planning on staying comfortable measuring old metrics, you’re not preparing for the real shift ahead.

 

 

What businesses should do today?

Embrace the grey area and awareness of new KPIs and give yourself the best chance of working towards them in the future. Don’t get lost in the GEO/AEO/XEO alphabet soup, but focus on some key actions: 

• Nail the SEO fundamentals: Accuracy in content titles, clear bullets, strong quality of content outputs.
• Layer in top-performing context: Reviews, FAQs and shopper discussion away from the online and social retail experiences are now essential optimisation fuel, so you need to prioritise those.
• Get to know your shoppers more deeply (and regularly): Learn their loves, loathes, and how they truly interpret your brand and their needs.
• Be intentional with your brand voice: Voice, tone and signals are now fair game to index and influence consumers.
• Experiment regularly with the new tools: Ask compelling questions to understand how AI interprets your brand.
• Test and learn: Start benchmarking what ‘context’ signals might matter, even if dashboards don’t measure it yet. 

 

Action beats anxiety every time

Here’s the truth: AI doesn’t get bored, doesn’t skim, and doesn’t forget. Shoppers may move fast, and they used to reward brevity. But in the AI-driven eCommerce world, it’s context and truth that are going to win.

 

If you’re optimising for real human truths – what shoppers love, what they hate and what they actually say – AI will follow.

 

SEO is far from dead. It’s expanding, changing, growing; and the businesses that understand what it’s looking for will succeed. 

 

---

 

Jay Tanel is Strategy Director for Digital Commerce at global creative agency Marks 

 

Main image courtesy of iStockPhoto.com and Urupong