From Rankings to Recommendations: What AI Search Really Changes for Marketers

Marketers are getting AI search wrong in two very different ways.

Either they treat AEO and GEO as an SEO rebrand, or they assume AI search platforms operate completely independently from traditional search.

Both views miss the point.

AI search is connected to traditional SEO, but it changes where influence happens, how brands are interpreted, what content needs to prove, and how marketers should measure success.

For years, search has been built around a familiar model: rank, click, visit, convert. That model has not disappeared, but it is no longer enough. Search is moving from a list of results towards a layer of recommendations, summaries, comparisons and actions.

As Jellyfish noted following Google Marketing Live 2026, search is becoming an AI answer layer that recommends, not just retrieves. Brands are no longer competing simply to be found in a list. They are competing to be included, interpreted and ultimately recommended by the engine itself.

Trust forms before the website visit

Today’s user journeys are unpredictable and our website is rarely the first meaningful digital interaction a buyer has with your brand.

People move between search engines, social feeds, videos, reviews, forums, AI tools, comparison sites and brand websites in no fixed order. By the time someone lands on your site, they may already have formed a view of who you are, what you offer and whether you are credible.

Take a simple example. Someone asks an AI tool, “What is the best CRM for small business?” The answer they receive may be shaped by vendor websites, review platforms, Reddit threads, YouTube videos, expert blogs, comparison articles and third-party mentions. In that moment, trust is being formed before the user comes anywhere near the CRM provider’s website.

Likewise, Large Language Models build their understanding of your company from countless external sources, drawing connections you cannot fully control. Critically, AI responses are trusted. McKinsey research on AI search found that 44% of AI-powered search users say it is their primary and preferred source of insight, ahead of traditional search, brand websites and review sites.

That should make marketers pause.

If buyers are using AI-generated answers to narrow options, compare providers and make sense of a category, your brand is being assessed before your owned experience has a chance to do the work.

AI search does not behave like a traffic channel

This is why traditional SEO measurement starts to break down.

Organic search has usually been treated as a performance channel. We have been able to track rankings, traffic, conversions and revenue with a fairly direct line between visibility and action.

AI-led search breaks that neat line.

AI engines are not listings-based, and driving traffic to your website is not one of their objectives. In many cases, they synthesise your content so the user does not need to visit your site.

If you measure AEO only by sessions and pageviews, you may conclude that it is killing your business. In reality, it may be influencing the buyer before they search for your brand by name.

The question is not “Did AI search send us traffic?”

The better question is: “Did AI search visibility help create a positive, accurate and useful perception of our brand before the user took action?”

That action may still happen through search. It may happen as a branded query. It may happen as a direct visit. It may happen later, after a second round of research or a sales conversation.

AEO drives influence. The commercial value is not always the immediate click. Often, you are fighting for real estate in the AI output so that when the user is ready to act, your brand is already in the shortlist.

Ideally, it is the brand they search for by name.

Visibility is not the same as perception

Traditional SEO has focused heavily on visibility. Are we ranking? Are we present? Are we getting clicks?

In AEO/GEO visibility means being present in AI responses. Does the AI mention your brand amongst a list of competitors or cite your content. But in AI responses, how you are described is just as important as whether you appear.

Being included in an AI generated list of software vendors is a start. But if the model describes your tool as overly complex while calling your competitor user-friendly, that visibility actively works against you.

In AI search being talked about is only half the battle. Being talked about correctly and positively is where the commercial value starts.

Our Jellyfish study on AI-mediated brand discovery makes this risk tangible. Its Agent Shopper research found cases where brands appeared frequently in AI shopping tasks, but were framed around a much narrower set of attributes than the brand itself would choose.

That is why emerging AI strategies focus heavily on narrative control. It is not enough to appear. When your brand surfaces, models need to position your offer accurately.

Generic SEO content is becoming a liability

Publishing basic informational content used to be a reliable way to capture search traffic.

For years, brands have published 101-level guides, “what is” articles and top-five-tips posts because keyword research suggested there was traffic available. Often, the process was simple: look at what ranks, combine the same points, polish the language and publish.

That era is over.

Generic top-of-funnel SEO content is becoming a liability, not an asset. By continuing to publish commodity content, you train users, search engines and AI systems to view your brand as a source of low-value information.

That does not mean introductory content is dead. Some buyers still need clear explanations. Some topics still require accessible entry points.

The problem is content that adds nothing.

In the AI era, if your content does not have a clear point of view, original examples, named expertise, proprietary data or practical value for your target audience, you should question whether it deserves to be published at all.

Decision-led content is what AI search can actually use

The alternative to top-of-funnel fluff is content built around complex choices.

Decision-led content helps someone make a choice, rather than just explaining a concept.

Information-led content asks: what is this topic? Decision-led content asks: what should someone do with this information?

It focuses on trade-offs, buying criteria and real-world application, the areas where expertise and judgement still matter.

A generic article might explain cloud computing. A decision-led article might compare AWS and Azure for healthcare startups with HIPAA requirements. The second requires more than a definition. It needs buying criteria, trade-offs, risk factors, practical context and a point of view.

AI search is especially relevant when users ask complex questions.

Which provider is right for my team? What should I compare? What are the risks? What does good look like?

Content that answers those questions is more useful to buyers, and more useful as source material for AI systems.

That should change how marketers think about content. AEO/GEO is not about adding a few question headings and hoping the model picks them up. It is about making your expertise easier to understand, verify and apply.

The most important GEO pages may not be blog posts

The most valuable GEO assets are often not blog posts. They are the structured pages that help AI systems understand the facts of your business: who you are, what you offer, who you serve, how you are different and what proof supports your claims.

That means brands need to look beyond the resource hub.

About and leadership pages help explain who sits behind the business.

Product and service pages explain what you offer, who it is for, how it works, what is included and what makes it different.

Documentation and knowledge bases provide structured, unambiguous facts.

Comparison and alternatives pages help users and AI systems understand where you sit in the market.

Case studies, original research and named expert commentary demonstrate expertise and real-world application.

The key question is not whether you have published another guide. It is whether AI systems can understand your brand well enough to recommend it in the right moments.

Can an AI tool tell what you do? Can it identify your strongest use cases? Can it find proof that your claims are true?

Those questions matter more than the next piece of generic top-of-funnel content.

AI search is no longer SEO’s responsibility alone

Once brand understanding is shaped across the wider ecosystem, AI search performance can no longer sit with SEO alone.

That is why Jellyfish increasingly talks about GEM, or Generative Engine Marketing. The argument is that brands need to think bigger than organic AI visibility because paid, earned, influencer and owned activity all shape how AI systems understand, rank and recommend brands.

That is a very different ownership model from traditional SEO.

SEO teams still matter. They understand discoverability, structure, content quality, technical accessibility and user intent. But they cannot control the whole picture alone.

Brand teams need to make positioning clear enough to be repeated accurately.

PR teams need to think about the third-party sources AI systems may trust.

Social, web and video content teams need to be aligned to produce consistent content that aligns with your search strategy.

Paid media teams need to understand how campaigns influence demand, recognition and search behaviour.

Product and ecommerce teams need to make factual information easy to find, compare and verify.

AI search makes marketing more connected. Every public signal can contribute to how your brand is interpreted.

AI search measurement needs a brand mindset

AI-led search behaves more like a brand channel than a traditional performance channel.

The connection between visibility and action is harder to measure, so marketers need a framework that combines multiple sources of data.

Traffic from AI engines does exist and should be part of the mix, but only tells a small part of the story. AI visibility tools are also part of the mix, but they need to be treated with care. AI responses are variable. Different users, prompts, locations, models and contexts can produce different answers. No AI visibility dashboard can perfectly tell you what every individual user saw.

A more useful measurement framework brings together:

• brand mentions in AI responses
• sentiment and perception
• citation sources
• branded search demand
• direct traffic
• assisted conversions

Brand mentions help you build a picture of presence. Sentiment helps you understand whether the context is favourable. Branded and direct traffic help you see where AI-influenced discovery may be turning into action.

That last point matters.

If a user encounters your brand in AI search in a positive way, one likely outcome is that they search directly for your brand. They may revisit through a bookmark, a search suggestion or a branded query rather than clicking a citation inside the original AI answer.

As Tom Roach of Jellyfish noted in Marketing Week, brand communication must now build model availability alongside mental availability. Marketers need to know how they are positioned inside language models.

For larger brands, the next step is to understand how AI systems describe and recommend your brand across key category prompts. By combining tools like Share of Model with direct traffic and branded search metrics, you can start to build a directional view of your true AI influence.

The future of search is not another acronym

The answer is not to abandon SEO.

The answer is to double down on the parts of SEO that were always worth doing: genuinely helpful content, clear technical foundations, useful structure, real expertise, strong internal linking, accessible pages and a deep understanding of what users need at each stage of the journey.

At the same time, marketers need to recognise that AI-led search is no longer the responsibility of SEO alone. Paid, owned and earned media all influence how LLMs learn and speak about your brand and your offer.

Leadership needs to build a cohesive strategy that connects content investment, PR, social media, and technical SEO. Winning in an AI search environment requires every team to understand their role in shaping the final recommendation.

So the future of search is not about chasing another acronym.

It is about making sure that when AI systems summarise your market, your brand is not only visible, it is understood, trusted and recommended.