Generative Search Is Changing SEO. Here Is What Shifts

Generative search changes the shape of SEO more than it changes the fundamentals. The goal remains the same: earn enough trust with search engines that they surface your content when someone is looking for what you offer. What changes is where that surfacing happens, what form it takes, and how you measure whether it is working at all.

AI Overviews, ChatGPT search, Perplexity, and whatever comes next are not replacing search intent. They are changing how intent gets satisfied. That is a meaningful distinction, and it is one that most SEO commentary is getting wrong.

What Generative Search Actually Does to the Search Results Page

The search results page has been changing for years. Featured snippets, knowledge panels, People Also Ask boxes, local packs, shopping carousels. Each of these reduced the dominance of the ten blue links without killing organic search. AI Overviews are a more aggressive version of the same pattern.

What generative AI adds is synthesis. Rather than pulling a single featured snippet from one authoritative source, an AI Overview can combine information from multiple sources into a single answer. That answer appears at the top of the page, often before any organic results. For informational queries, particularly the kind that used to reliably drive top-of-funnel traffic, this is a significant change.

The traffic impact is real but uneven. High-volume informational queries, the “what is”, “how does”, “difference between” type searches, are most exposed. Commercial and transactional queries, where someone is ready to buy or compare, are less affected. Google has a commercial incentive to keep those clicks happening. It is not in their interest to let generative AI answer “best project management software for a 50-person team” without sending anyone to a comparison page or a paid result.

I have managed paid search and SEO programmes across 30 industries over two decades, and the pattern I keep seeing is that the parts of organic search most at risk are the ones that were always the weakest from a commercial standpoint. Traffic without intent. Informational content that attracted visitors who were never going to convert. If generative AI absorbs some of that traffic, the honest question is how much it actually mattered.

How Do You Get Cited in AI-Generated Answers?

This is the question every SEO team is asking right now, and the honest answer is that the signals look familiar. AI systems, whether Google’s Overview feature or third-party tools like Perplexity, tend to cite sources that already rank well in traditional search. The correlation is not perfect, but it is strong enough to be directionally useful.

What earns citations tends to be content that is specific, clearly structured, and written by someone with demonstrable expertise on the topic. Vague, hedged, keyword-stuffed content that ranked on pattern-matching alone is not what AI systems are pulling from. They are pulling from content that actually answers the question.

There is useful framing from Moz’s work on generative AI and content on what makes content viable in this new landscape: authority signals, structured answers, and demonstrable topical depth. None of that is new. What is new is that the bar for “good enough” has risen, because AI can now synthesise answers from multiple sources and your content only gets cited if it contributes something distinctive.

Practically, that means a few things. First, answer questions directly and early in your content, not after three paragraphs of preamble. Second, use clear structure, headers that match the questions people are actually asking, not the questions that sound impressive. Third, demonstrate expertise in a way that is verifiable, named authors with real credentials, citations to primary sources, data that is actually yours rather than recycled from someone else’s post.

This is also where brand signals start to matter more than they did in traditional SEO. If your brand is mentioned in forums, reviews, industry publications, and third-party sources, AI systems have more reason to treat you as a credible source. That is a longer game than keyword optimisation, but it is a more durable one.

If you are working through your broader SEO approach alongside these changes, the Complete SEO Strategy hub covers the full picture, from technical foundations to content architecture to how performance measurement needs to evolve.

Does E-E-A-T Mean More Now Than It Did Before?

Experience, Expertise, Authoritativeness, Trustworthiness. Google’s E-E-A-T framework has been part of the quality rater guidelines for years, but it has always felt slightly abstract when translated into actual SEO practice. Generative search makes it more concrete.

When an AI system is deciding which sources to cite, it is effectively making an E-E-A-T judgement. Does this source know what it is talking about? Is there evidence that the author has real experience with this topic? Is the site trustworthy enough to be cited in an answer that will be seen by millions of people? These are not algorithmic questions in the traditional sense. They are closer to editorial judgements.

I judged the Effie Awards, which are awarded for marketing effectiveness rather than creativity alone. One of the things that became clear sitting on that panel was how rarely brands could actually demonstrate the causal link between their marketing activity and their business outcomes. Most entries were good at showing correlation. Very few could show mechanism. E-E-A-T in content is the same challenge: it is easy to claim expertise, much harder to demonstrate it in a way that is independently verifiable.

The practical implication is that content strategy needs to be more tightly connected to genuine organisational knowledge. If you are publishing content about a topic your team does not actually understand deeply, that gap is more visible now than it was when SEO was primarily a pattern-matching exercise. Thin content with good backlinks could rank. It is much less likely to be cited in an AI answer.

What Happens to Keyword Strategy in a Generative World?

Keyword research does not disappear, but its role shifts. Traditionally, keyword research was primarily about finding search volume and ranking opportunity. In generative search, the more useful frame is understanding the questions that people are actually trying to answer, and whether your content provides a better answer than what currently exists.

The distinction matters because AI systems are not matching keywords to pages. They are matching questions to answers. A page that ranks for “content management systems and SEO” because it contains those words is not the same as a page that genuinely explains how CMS choices affect search performance in a way that someone could act on. The latter is far more likely to earn a citation.

Long-tail queries become more interesting in this context, not less. Generative AI handles generic informational queries efficiently. What it handles less well is nuanced, specific, or experience-based questions where the answer depends on context. “What is content marketing” is a question AI can answer without citing anyone in particular. “How do you rebuild an SEO programme after a site migration goes wrong” is a question where specific, experienced voices are more likely to be surfaced.

This is not a new insight exactly. Understanding search behaviour and how it connects to conversion has always been more valuable than chasing volume for its own sake. Generative search makes that principle more urgent.

How Does Measurement Change When Clicks Decline?

This is the part of the generative search conversation that makes me most cautious about confident predictions. Measurement in SEO has always been messier than people admit. I have spent years working with GA, GA4, Adobe Analytics, and Search Console across large programmes, and the honest truth is that these tools give you a perspective on what is happening, not an accurate count. Referrer loss, bot traffic, classification issues, dark social, implementation quirks. The gap between “what the tool reports” and “what actually happened” has always been wider than most teams acknowledge.

Generative search makes this harder. When someone reads an AI Overview and then types your brand name directly into their browser, that visit appears as direct traffic. When someone sees your brand cited in a Perplexity answer and searches for you by name, that appears as branded organic. Neither of these gets attributed to the AI citation that prompted the behaviour. The influence is real, the attribution is invisible.

The response to this is not to find a better attribution model. It is to measure a wider set of signals and to be honest about what you can and cannot know. Brand search volume trends. Direct traffic trends. Share of voice in your category. Referral traffic from AI tools where it is trackable. Conversion rate from organic visits that do arrive. None of these individually tells you the full story, but together they give you a directional read on whether your visibility in generative search is growing or shrinking.

The teams that will get this wrong are the ones insisting on last-click precision in a world where the click is increasingly not the primary signal. I saw this pattern play out in paid search years ago, when attribution modelling became an obsession that often produced more confidence than accuracy. Honest approximation is more useful than false precision.

The HubSpot research on consumer trends and the future of SEO points toward the same conclusion: the metrics that matter are shifting toward brand engagement and intent signals, not just organic session counts.

What Does This Mean for Link Building and Domain Authority?

Links remain a signal. The question is whether they remain the dominant signal they have been for the last two decades, and the honest answer is probably not in isolation.

Traditional link building was effective partly because it was a proxy for authority. If many credible sites link to you, that is evidence that your content is worth referencing. AI systems care about the same underlying thing, genuine authority, but they have more ways to assess it. Brand mentions without links, citations in forums and communities, author reputation signals, structured data that makes expertise verifiable. These are all inputs that matter in a generative search environment.

That does not mean links are irrelevant. Competitive link research still reveals meaningful gaps in authority relative to your competitors, and a strong backlink profile remains a useful foundation. What changes is that link building alone, without genuine content quality and brand signals, is less likely to produce durable results in a world where AI systems are making more nuanced authority assessments.

The practices that were always slightly dubious, bulk link acquisition, low-quality directory submissions, exact-match anchor text manipulation, are more exposed now. Not because Google has suddenly become better at detecting them, though it has, but because the question AI systems are asking is different. They are not asking “how many links does this page have.” They are asking “is this a source worth citing.”

Which Businesses Are Most Exposed and Which Are Most Protected?

Not every business is equally affected by generative search. Understanding where you sit on that spectrum is more useful than applying generic advice.

The most exposed businesses are those whose organic traffic is primarily informational and whose content does not represent genuine proprietary expertise. Publishers that built large content libraries on generic topics, aggregators that ranked on volume rather than depth, sites whose SEO strategy was primarily about capturing top-of-funnel informational queries. These businesses are seeing real traffic pressure already.

The most protected businesses are those with genuine brand authority, specific expertise that AI cannot easily synthesise, or strong commercial intent in their query mix. A B2B software company with deep technical documentation and a strong brand in its category is less exposed than a generic blog covering the same topics. A retailer with strong branded search and a clear transactional intent profile is less exposed than a content site living on informational traffic.

There is also a structural advantage for businesses that have diversified their acquisition channels. I ran a performance marketing programme at lastminute.com where a single well-executed paid search campaign generated six figures of revenue in roughly a day from a music festival launch. The speed was remarkable, but what made it sustainable was that paid search was one channel among several, not the whole strategy. Organic search works the same way. Businesses that treat SEO as one part of a broader acquisition picture are less exposed to any single change in how search works.

The foundational SEO principles that have always separated durable programmes from fragile ones, genuine expertise, clean technical foundations, content that serves real intent, are the same ones that determine resilience in generative search.

What Should SEO Teams Actually Do Differently?

The temptation with any significant shift in search is to treat it as a new set of tactics to master. Add structured data for AI citations. Optimise for conversational queries. Target “generative engine optimisation.” Some of this is useful framing. Most of it is the same pattern the industry runs every time something changes: rebrand the fundamentals, sell urgency, charge for workshops.

What SEO teams should actually do is more straightforward, and more demanding. Audit your content honestly. Which pieces represent genuine expertise? Which were written to rank rather than to inform? The latter category is increasingly vulnerable, and the time spent defending it is time not spent building something more durable.

Invest in author credibility. Named authors with verifiable expertise, author pages that establish credentials, content that reflects real organisational knowledge rather than synthesised generalities. This matters for E-E-A-T signals and it matters for AI citation likelihood.

Build brand signals deliberately. Coverage in industry publications, mentions in communities where your audience is active, partnerships that generate genuine third-party references. These are harder to manufacture than links, which is precisely why they are more valuable.

And measure differently. Stop optimising for metrics that will become less meaningful as generative search grows. Start tracking brand search volume, direct traffic trends, and share of voice alongside traditional organic metrics. The picture will be incomplete, but it will be more honest.

There is a broader point worth making here. The businesses that are anxious about generative search are often the ones that built their SEO on patterns rather than principles. If your organic visibility depends on algorithm behaviour that could change, you are not in a strong position regardless of what search looks like next year. The ones with genuine expertise, real brand equity, and content that serves actual user needs are in a much better position, not because they have cracked the generative search code, but because they built something worth citing in the first place.

If you are rethinking your SEO approach in light of these changes, the Complete SEO Strategy hub at The Marketing Juice covers the strategic and tactical foundations that hold up regardless of how search technology evolves.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions