Okay, I need to talk about something that's been rattling around my brain since last week's piece. And no, it's not the existential dread of watching Google slowly eat its own product (though that's also happening). It's the question that everyone asked after I/O and almost nobody answered well: what do you actually do now?

Because here's the thing. Last week I laid out what Google announced—generative UI building tools inside Search, information agents that read your content and deliver summaries without sending a single visitor, AI Mode queries running three times longer than traditional searches. The internet did what the internet always does: half of it screamed "SEO is dead" and the other half insisted nothing has changed. Both sides were wrong, which is becoming a depressingly reliable pattern.

The real situation is more uncomfortable than either camp wants to admit. Google didn't remove web results. It just made them… optional. And if you've ever been "optional" in a relationship, you know how that story ends. (Not great. The answer is not great.)

Traditional SEO ran on a beautifully simple chain: create content that matches search intent, earn a high ranking, convert the resulting traffic. Three links. Clean. Elegant. The kind of thing you could explain to a client over a single coffee.

Here's what broke: the first link still works. Google confirmed—both in its pre-I/O optimisation guide and in a clarification posted the day after the keynote—that AI features depend on the same ranking systems that power organic results. Relevance, authority, well-structured content. The fundamentals are intact. If you've been doing good SEO, congratulations, that work wasn't wasted. (I can feel the collective exhale from here.)

But the second link? The ranking-to-traffic connection? That's where things get uncomfortable. Position one on a query where AI Overviews appear now delivers a click-through rate of roughly 1.6%. It used to be 7.3%. Organic CTR on affected queries has fallen between 38% and 61%, depending on whose research you trust. And those numbers were measured before I/O announced generative UI, mini apps, and always-on information agents.

So ranking is now necessary but insufficient. You need it to be eligible for AI citation. But ranking alone no longer reliably converts into traffic. The click—that beautiful, trackable, monetisable click—is being quietly compressed out of an expanding category of queries.

This is the shift. Not the death of SEO. The expansion of what SEO has to mean.

I know, I know. "Visibility without clicks" sounds like "a restaurant without food." But hear me out, because this is where the strategy actually lives.

First: citation. When Google's AI Mode builds a comparison, surfaces a recommendation, or constructs an overview, it cites sources. Those "According to [Your Brand]" references deliver brand exposure, authority signalling, and sometimes even a click—just at a lower rate than we're used to. Research from Seer Interactive found that brands cited within AI Overviews earned 35% more organic clicks and 91% more paid clicks than brands that weren't cited. Citation isn't a consolation prize. It's becoming the main mechanism through which search delivers brand value. Whether you're a SaaS company, a law firm, a local plumber, or a B2B manufacturer—if you're not being cited, you're functionally invisible.

Second: data contribution. When an information agent monitors your product pages, or generative UI builds a comparison tool using your specifications, your content is powering the experience—even if no one sees your URL. This is harder to measure (read: currently almost impossible to measure), but the brands whose data is comprehensive enough to be selected by AI systems are the ones showing up inside the experience. The brands whose data is incomplete or poorly structured are just… absent. Like they never existed. Which is a fun thing to contemplate at 2 AM. (I don't recommend it.)

Third: brand query generation. Amsive's research found that branded queries with AI Overviews see an 18% increase in CTR, while generic queries see a 34% to 46% decrease. Read that again. The most durable form of visibility in AI-mediated search isn't ranking for generic terms. It's being the brand users search for by name after the AI has done the initial discovery. Brand investment isn't separate from search strategy. It's upstream of it.

Here's where I need to get a bit nerdy. (More nerdy. I'm aware I'm already operating at concerning levels of nerd.)

AI search systems use a process called query fan-out. A single user prompt gets decomposed into multiple sub-queries that run simultaneously. Someone typing "best project management software for remote teams under 50 people" doesn't trigger one search. It fans out into sub-queries covering pricing tiers, integrations, mobile app quality, user reviews, security features, scalability, customer support, and comparison articles. The AI assembles its response from the combined results across all those parallel searches.

This fundamentally changes what "optimised" means. A page that ranks well for one keyword but only addresses one dimension of the topic will perform poorly across the fan-out. A site that covers the topic comprehensively—across multiple pages, with genuine depth in each—accumulates citation weight through something called reciprocal rank fusion. Sources that appear consistently across multiple sub-query results earn higher composite scores.

Google's own optimisation guide reinforces this. It states that Google evaluates "how consistently a topic is covered across a website." That's not a ranking factor in the traditional sense. It's a topical authority signal. And it directly maps onto how fan-out retrieval works.

The practical consequence? The unit of optimisation in AI search isn't the page. It's the topic. If your content strategy is still built around keyword research tools and individual page targets—whether you're a SaaS company mapping competitor keywords or a local accountancy firm chasing "tax advice [city name]"—this is the structural adjustment that matters most.

Audit your content for commodity versus non-commodity status. Google's guide explicitly distinguishes between commodity content—information that restates what's available elsewhere—and non-commodity content offering original data, primary research, or genuine expertise. Non-commodity content is what AI systems must cite rather than merely summarise. For every piece in your content library, ask: does this contain information an AI couldn't assemble from other sources? If not, that content is the most exposed to being synthesised without a click. If yes, it has structural protection—but only if it's well-structured enough for AI retrieval systems to find and extract the unique bits.

Restructure around topic clusters, not keyword lists. Map your priority topics. Identify the sub-queries AI systems are likely to generate through fan-out. Evaluate whether your site addresses each dimension with genuine depth. Gaps in coverage are gaps in citation eligibility. If you're a cybersecurity firm covering "endpoint protection" but have nothing on compliance frameworks or incident response integration, you're invisible to those sub-queries. Your composite score suffers accordingly.

Treat structured data as non-negotiable infrastructure. Schema markup, clean data feeds, machine-readable content attributes—these are how AI systems identify, categorise, and retrieve your content during fan-out. Roughly 70% of websites still haven't implemented schema markup. If your content is locked in unstructured prose, you're functionally invisible to the systems now mediating an expanding share of discovery. This applies to every sector: product catalogues, service listings, professional credentials, local business information. All of it.

Invest in brand as a search strategy. The Amsive data isn't ambiguous. Branded queries outperform in AI-mediated search. Every activity that increases the likelihood that someone types your brand name—or that an AI system recognises you as a topical authority—is now a search activity. Content marketing, PR, community building. Analysis of over one million links cited across ChatGPT, Claude, Gemini, and Perplexity found that roughly 94% of citations come from non-paid sources, with earned media accounting for 82% of all citations. Getting your brand covered by reputable publications isn't just a PR exercise. It's citation acquisition. (Which is a phrase that didn't exist two years ago, and now I'm using it unironically. The future is weird.)

Build measurement infrastructure now. Google hasn't provided Search Console filters to differentiate AI Mode traffic from standard organic traffic. Information agents may consume your content without it appearing in analytics at all. The best available approach combines imperfect signals: monitoring impression-to-click ratio shifts in Search Console, tracking branded search volume as a proxy for AI-driven awareness, analysing server logs for AI crawler activity, and using Bing's AI Performance dashboard for citation-level data. None of this is clean. But making strategic decisions with no measurement at all is worse.

Here's the part that genuinely keeps me up at night. Google's optimisation guide says AEO and GEO are "still SEO." The Search team advises focusing on current needs. The official messaging is continuity and calm.

But the same week, Google announced the biggest upgrade to the Search box in over 25 years. AI Mode surpassed one billion monthly users. They launched always-on information agents. They showed generative UI building custom tools inside search results. User behaviour is already shifting toward longer, multimodal, follow-up-heavy queries.

The infrastructure is changing faster than the guidance. That gap is the single most important thing to understand about where search is heading.

The fundamentals still apply. But the environment they apply within now demands a broader, more strategic approach than keyword rankings and position tracking alone. Whether you're a SaaS startup, a regional law firm, a publisher, or a local services business—the shift from traffic acquisition to citation and visibility strategy isn't a future possibility. It's the current reality. I/O just made it harder to pretend otherwise.

The good news? If you've been building genuine expertise, creating non-commodity content, and investing in topical authority, you're already better positioned than most. The bad news? "Most" is a very low bar right now.

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.

😍It was great, all I needed

🙂Good, not great

🙄It sucked