There is a number that every ecommerce business owner needs to understand right now: 93%.

That is the percentage of Google AI Mode searches that end without a single click to any website. Not 30%. Not 50%. Ninety-three.

For context, AI Overviews — the AI summaries that have been appearing in standard Google search results since 2024 — already suppress organic click-through rates by 58 to 61%. Those felt alarming when the data came out. AI Mode makes them look mild.

We are not, to be clear, in a period of incremental change. We are in the middle of a structural rearchitecting of how Google works — and for ecommerce brands, the window to respond is now, not next quarter.

What Is AI Mode, and Why Does It Change Everything?

Google AI Mode is an end-to-end AI search experience powered by Gemini 3 (the most recent version, rolled out globally in January 2026). When a user activates AI Mode, they do not see the traditional ten blue links. They see a conversational AI response — synthesised, structured, and cited — with no guarantee that any specific brand appears, and no guarantee that a user ever leaves Google to find out more.

It works differently from AI Overviews. Where AI Overviews appear alongside standard search results, AI Mode replaces them. And where AI Overviews issue a single query, AI Mode fans out — running up to 16 simultaneous sub-queries to build a comprehensive answer. The result is more thorough, more confident, and less dependent on any single source.

For SEO professionals, one data point stands out above all others: only 13.7% of citations overlap between AI Overviews and AI Mode. They draw from fundamentally different content ecosystems. A brand that appears in AI Overviews is not automatically visible in AI Mode. Both surfaces require independent optimisation — and most brands have not meaningfully optimised for either.

The Zero-Click Reality Is Not Coming. It Is Here.

Consider where we are as of early 2026:

  • AI Overviews now appear on approximately 48% of all Google searches, up from 31% twelve months ago.
  • On informational shopping queries — “best running shoe under £100”, “which protein powder has least sugar” — 83% now trigger an AI Overview.
  • 60% of all Google searches end without a click to any external website.
  • Organic CTR for queries with an AI Overview has dropped from 1.76% to 0.61% — a 65% reduction.

These are not projections. They are measurements from the current search landscape, drawn from Semrush, Ahrefs, Seer Interactive, and Pew Research Center data published between late 2025 and early 2026.

For ecommerce, this bifurcates the search journey into two distinct phases. The research phase — “what’s the best,” “how do I choose,” “what works for” — is now overwhelmingly handled by AI. The purchase phase — specific product searches, brand queries, direct intent — still carries more traditional search behaviour, though AI is encroaching there too.

The implication is direct: if your brand is not cited in AI-generated answers during the research phase, you are not part of the consideration set. You do not lose a click. You are never seen at all.

Google Has Also Changed What Happens After the Search

AI Mode is not just changing how people search. It is changing what Google does with the result.

Checkout in AI Mode — launched in early 2026 — allows Google to complete purchases inside its own interface. A user asks “where can I buy a waterproof jacket under £150” and, without ever visiting a brand’s website, can complete a transaction through Google. This is not a future possibility. It is live infrastructure, built on the Universal Commerce Protocol (UCP) that Google launched on January 11, 2026 in partnership with Shopify, Wayfair, Target, and others.

The consequence for brands is significant. If a sale is completed inside Google’s ecosystem, no visit is recorded. The brand ships an order. Google earns the commercial relationship.

Understanding this shift is not optional for brands with serious ecommerce ambitions. Adapting to it — by ensuring your brand is present at the AI citation layer, not just the click layer — is now the core work of search optimisation.

Google Merchant Center Is Now Your Most Important SEO Asset

Here is what often surprises ecommerce clients when we first have this conversation: the brands most likely to appear in AI Mode product recommendations are not necessarily those with the best-optimised websites. They are the ones with the most complete, attribute-rich product data in Google Merchant Center.

This is the core of how the AI shopping layer works. Google’s AI Mode, AI Overviews, and Gemini all draw heavily on structured product data — GMC feeds, schema markup, and entity signals — rather than crawling websites in the traditional sense. A product with a full GMC entry (brand, GTIN, accurate descriptions, lifestyle imagery, product highlights, certifications, provenance signals) is far more legible to these systems than a well-written product page with a thin or incomplete feed.

There is also a new GMC feature worth understanding specifically: Business Agent. Currently rolling out in the US (with a simpler Q&A version already live in the UK, US, Canada, Australia, and India), Business Agent adds a “Chat” button to brand profiles in Google Search. Shoppers can ask questions about products — “Is this dishwasher safe?”, “What’s the thread count on this?”, “How do I know which size to order?” — and receive answers generated by Gemini directly from your Merchant Center data.

The quality of those answers is entirely determined by the quality of your feed. A bare-minimum GMC entry produces bare-minimum answers, or no answers at all. A rich, complete feed means your brand can answer customer questions at scale — without staff, without a chatbot, inside the search result.

What the March 2026 Core Update Tells Us

If you have seen organic traffic drops on any of your brand’s websites in the last few weeks, you are not alone. Google’s March 2026 Core Update began rolling out on February 24, with Semrush’s volatility sensor peaking at 9.5 out of 10 — among the highest readings of the year so far.

The update follows a clear pattern that has been consistent across Google’s core updates since mid-2024: sites with strong brand signals, genuine expertise, and content that earns AI citations are recovering or growing; sites with thin content, weak entity signals, and poor structured data are declining.

This is not algorithm capriciousness. It is the logical consequence of Google reorienting its ranking systems toward the same signals that its AI surfaces prefer. E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is no longer an abstract quality guideline. It is the operating principle of both traditional rankings and AI citation eligibility.

What Ecommerce Brands Should Be Doing Now

The brands that will perform best through the rest of 2026 and into 2027 share a common approach. They are not treating AI search as a future problem. They are treating feed quality, entity signals, and AI citation as the primary work of search optimisation today.

Specifically, that means:

1. Audit your GMC feed for AI-readiness, not just compliance.
Most brands have feeds that meet Google’s minimum requirements. Very few have feeds that are optimised for the AI layer — meaning rich descriptions (1,000–1,500 characters, not 200), product highlights, accurate categorisation, lifestyle imagery, certifications, and provenance detail. This is where the gap between visible and invisible in AI search is decided.

2. Build entity signals, not just backlinks.
AI search engines cite brands with strong entity presence — consistent information across third-party directories, press coverage, structured data that confirms who you are and what you sell. A backlink from a high-DR site matters less than consistent, accurate brand information appearing across multiple trusted sources.

3. Create content that AI engines want to cite.
Content that performs in AI Overviews and AI Mode tends to be specific, authoritative, factual, and structured around clear questions and direct answers. Buying guides, detailed FAQs, provenance stories, and comparison content — where the brand has genuine expertise and is not afraid to be definitive — outperform generic category pages.

4. Track AI referral traffic as a metric.
From 2026 onward, traffic from ChatGPT, Perplexity, and Google AI Mode is trackable in Google Analytics as referral traffic. If you are not monitoring it, you have no visibility into how much of your discovery is now happening in AI channels — and no baseline against which to measure optimisation gains.

5. Treat GMC as a publishing platform, not a feed submission tool.
The brands winning the AI shopping layer are the ones updating their product data with the same rigour they apply to their website copy. New products, seasonal updates, certifications, promotional events — all of it should flow through a well-maintained, attribute-rich GMC feed.

This Is Not the End of SEO. It Is the Beginning of a More Demanding Version of It.

We want to be precise about what we are describing — because the “SEO is dead” narrative is as wrong now as it has been every time it has been declared.

Google still processes billions of searches every day. Organic search still drives substantial ecommerce revenue. Core Web Vitals, technical SEO, and well-structured product pages still matter. What has changed is the sufficient condition for visibility. Being well-optimised for traditional search is now necessary but no longer enough.

The additional requirement — structured data, entity authority, feed completeness, answer-optimised content — is not separate from SEO. It is SEO, expanded into the channels that are now doing the recommendation work that organic results used to handle.

The brands investing in this layer now are building an advantage that will compound. The brands waiting for the landscape to settle before they act are ceding ground that will become progressively harder to recover.

How Consilium Design Approaches This

We work with UK ecommerce brands across food and drink, skincare, homewares, and premium lifestyle — categories where provenance, story, and product quality matter, but where discovery increasingly happens in AI-mediated environments that most brands have not yet considered.

Our approach combines technical feed optimisation, entity-building strategy, and content architecture designed for AI citation — all anchored in the commercial reality of what actually drives revenue, not just visibility.

If your brand is seeing unexplained traffic declines, if your GMC feed has not been reviewed for AI-readiness, or if you are not yet tracking how your products appear in AI-generated search answers, we would be glad to have that conversation.

Sources: Semrush, Ahrefs, Seer Interactive, Pew Research Center, Search Engine Land, Google Merchant Center Help, BrightEdge, Visibility Labs analysis of 20.9M shopping keywords (early 2026).