Search Without Clicks: What Google I/O 2026 Means for SaaS and Ecommerce

Sundar Pichai opened the I/O 2026 keynote with a familiar refrain: Google has been "AI-first" since 2016. A decade-long head start that, by his own framing, has now produced "the biggest reinvention of the search box in 25 years." The pre-2022 version of Pichai might have called that urgency. The post-ChatGPT version of him is calling it a strategy.

Whatever you call it, the change is real. On May 19, Google announced the most ambitious overhaul of its search product since the link became the unit of the internet. And buried in the announcement is a quiet thesis: the link is no longer the point.

This post does two things. First, it summarizes what Google actually shipped. Second, it walks through what the data says, and how the implications split between Services / SaaS businesses on one side, and ecommerce on the other. Five charts at the bottom you can drop into a deck or post to LinkedIn.

Google's own blog is the primary source. Where we cite third-party data, we link to it.

What Google actually announced

The Search-related announcements at I/O 2026 fall into five buckets.

A new search box. Google calls it "the biggest upgrade to the Search box in over 25 years." It expands to fit longer queries, accepts text, images, files, videos and Chrome tabs as inputs, and uses AI to suggest follow-up questions rather than just autocomplete keywords. It rolled out yesterday in every country where AI Mode is available.

Gemini 3.5 Flash powers AI Mode globally. AI Mode now runs on Gemini 3.5 Flash for every user. Google claims Flash is four times faster than other frontier models, and twelve times faster inside its agentic coding platform Antigravity. AI Mode has surpassed one billion monthly active users, with queries more than doubling every quarter since launch.

Information agents. Personalized agents that run in the background 24/7, scanning the open web plus Google's own real-time data for changes that match criteria you set. Apartment hunting, sneaker drops, sector news. Launches this summer for Google AI Pro and Ultra subscribers.

Generative UI and mini apps. Powered by Antigravity, Search now builds custom layouts, interactive visuals, dashboards and trackers on the fly inside the search results page. Generative UI is free for everyone this summer. Persistent "mini apps" arrive for paid subscribers in the months after.

The Universal Cart. Worth its own announcement. Google launched an intelligent cart that follows you across Search, Gemini, YouTube and Gmail, built on a new Universal Commerce Protocol (UCP) and Agent Payments Protocol (AP2). Launch partners include Nike, Sephora, Target, Ulta, Walmart, Wayfair and Shopify merchants such as Fenty and Steve Madden.

The headline numbers Google chose to disclose are unflinching. AI Overviews now reach 2.5 billion monthly active users. The Gemini app went from 400 million to 900 million MAU in a year. Daily Gemini requests grew sevenfold. Token throughput across Google's surfaces went from 9.7 trillion per month in 2024, to 480 trillion last year, to 3.2 quadrillion this year. Capex sits at $180 to $190 billion, six times what it was four years ago.

Those are the numbers Google wants you to see. The interesting ones are the ones it does not.

The data picture: a click economy in retreat

Ahrefs found that for queries where Google now shows an AI Overview, the position-one click-through rate drops from 7.3% to 1.6%. A 78% cut. Seer Interactive's study of 25.1 million impressions found that 93% of Google AI Mode sessions end with zero clicks to any external site. Across all US Google searches, 58.5% now end with no click.

Press Gazette's analysis of Chartbeat data across 2,500 publisher sites showed global search referrals fell 33% in 2025, with US-specific declines of 38%. The Reuters Institute reports that media executives expect a further 43% average decline over the next three years.

Pichai told the audience that "when people use our AI-powered features in Search, they use Search more." The data confirms that. What he did not say is that the additional searching does not always come with a link out. The retention rate on AI Mode is the headline business-model story of the next two years.

But Google is no longer the only relevant search engine, and ChatGPT is no longer the only relevant AI search engine.

Similarweb's January 2026 Gen AI Tracker shows ChatGPT held 86.7% of Gen AI website traffic share a year ago. Today it holds 64.5%. Gemini went from 5.7% to 21.5% in the same window. DeepSeek entered the market at 4.2%. Grok crossed 3%. The single-platform AI search era ended sometime around Q3 last year. Most marketing dashboards have not noticed yet.

There is one more measurement quirk worth knowing before we split into segments. Loamly analyzed 446,405 AI-referred visits and found that 70.6% of them arrived without referrer headers, misclassified as "direct" in Google Analytics. The dark AI traffic converts at 10.21%, four times the rate of standard direct traffic. So almost everything below understates the channel, sometimes by a lot.

What it means for Services and SaaS

Most B2B service and SaaS businesses live or die on the informational query. How do I, best tool for, compare X and Y, is X worth it. These are exactly the queries where Google's AI Overview lives. Top of the funnel, where the user is still learning, where you used to win the click by ranking for the right keyword and writing useful content.

The math has changed in two ways.

One: volume

Seer Interactive found that organic click-through rates dropped 61% for queries where AI Overviews appeared. For B2B technology specifically, AI Overview exposure runs around 70% of the query types your buyers use. 73% of B2B websites lost traffic between 2024 and 2025, averaging a 34% decline year over year. CTRs for B2B software categories fell up to 30%.

If you sell software through a content-led SEO motion, half of your funnel just thinned out.

Two: intent

AI-referred users sign up at 1.66%, eleven times the rate of Google organic. ChatGPT-referred users convert at 14.2% on transactional sites versus Google's 1.76%. Claude, despite a 2% market share, converts highest of any AI source at 16.8%. Perplexity, where 30% of the user base is senior leadership, sits at 10.5%. Ahrefs reported that AI traffic represented 0.5% of its visitors but drove 12.1% more signups than the rest combined.

The shape is consistent across studies. AI search is a smaller, slower-flowing river. It is also higher in altitude. Users arrive having already done the comparison, the synthesis, the "is this even the right category" thinking. They click because they need to transact. They are upstream of organic intent, not downstream.

AI visibility is becoming a multi-platform, low-volume, high-intent channel. Plan accordingly.

What this means in practice for Services and SaaS:

The keyword-rank-CTR playbook is no longer the unit of growth. The unit is being cited as the answer. Whether the citation ships traffic or not, it shapes what the next user types into Google directly, what they ask their colleague, what they remember when budget approval comes around. Pichai said "people love AI Overviews and they're coming back to Search more." The downstream effect for any individual SaaS company is that brand becomes the only durable asset. Generic positioning gets buried inside the AI Overview. Branded queries punch through it.

The good news for SaaS is that IT and B2B SaaS already receives the highest AI referral rate of any measured vertical, at 2.8% of all visits, with leading Loamly cohort members seeing 5 to 15% of total traffic from AI sources. The category is structurally well suited to AI citation: product comparisons, integration questions and developer documentation are exactly the content types LLMs use to synthesize answers.

The bad news is that two thirds of the AI traffic arriving is invisible in default Google Analytics. If you have not audited for dark AI traffic, you are budgeting blind.

What it means for ecommerce

Ecommerce got a different I/O announcement, and the shape of the implications is different.

Google's Universal Cart takes the agentic commerce stack from interesting to operational. The cart follows you across Search, Gemini, YouTube and Gmail. It runs price tracking, stock alerts, payment perk optimization and product compatibility checks in the background. It checks out through Google Pay or hands you off to the merchant. UCP launch partners include Nike, Sephora, Target, Ulta, Walmart, Wayfair, and Shopify merchants like Fenty and Steve Madden. Agent Payments Protocol (AP2) lets users set spending limits and brand criteria, then have an agent transact within those guardrails.

What this turns into operationally: product discovery moves up the stack, and checkout becomes a programmable layer. The merchant of record stays the merchant. The discovery layer stops being the merchant's site.

Travel and hospitality saw AI referral traffic grow 1,700% between July 2024 and February 2025. Retail and ecommerce grew 1,200%. Financial services, 1,200%. These are the categories where synthesis matters most (which hotel for our anniversary in Lisbon, what's the best 65-inch OLED under $1,500, should I take this mortgage rate) and where the AI can do the comparison up front but cannot complete the transaction.

35% of US consumers now use AI at the product discovery stage. 13.6% use traditional search for the same thing. The discovery layer has effectively moved.

The conversion data confirms this is not just exposure. Target receives roughly 15% of its referral traffic from ChatGPT alone. Etsy gets over 20%. On transactional sites, ChatGPT-referred buyers convert at 7% versus Google's 5%.

What changes operationally for ecommerce:

Product data quality stops being optional. Universal Cart, AP2 and every retrieval-augmented AI search depend on structured product information. The retailers that win the agentic commerce era are the ones whose product feeds are complete, accurate, and machine-readable. The ones that won SEO by writing good copy will not, by default, win this.

Brand mention share inside AI responses becomes the upstream marketing KPI. When ChatGPT recommends "the best running shoes for flat feet", the brands it names are the ones that get the conversion. The brands it does not name might as well not exist. CeraVe holds 27.17% of AI mention share in beauty. Apple holds 54.38% in electronics. Those are not accidents. They reflect deliberate structured-content investment.

[Insert Chart 05: AI search rewards the known brand. File: graph-5-branded-vs-generic.png]

Branded queries that trigger AI Overviews see CTR rise 18%. Generic queries that trigger AI Overviews see CTR fall 34 to 46%. Brands cited inside the AI Overview get 35% more organic clicks and 91% more paid clicks than uncited brands in the same SERP.

The AI search era does not make brand less important. It makes brand the only thing that still works.

Where this is heading

Two trajectories are worth watching.

First, platform fragmentation keeps going. ChatGPT lost 22 share points in 12 months. Gemini's distribution advantage through Android, Chrome and Workspace is just starting to compound. DeepSeek and Grok together hold 7%+ and rising. By Q3 2026, optimizing for ChatGPT alone will leave at least 35% of the AI traffic market unaddressed.

Second, the retention loop tightens. 93% of AI Mode sessions stay in the AI Mode pane. 75% never leave the interface at all. AI referral traffic peaked in July 2025 and is down 42.6% since. The traffic that escapes the retention loop is increasingly self-selected for high intent. This is why AI traffic converts. It is also why total volume will keep underwhelming.

The agentic commerce future Google sketched with Universal Cart is the first concrete answer to "what comes after the link." It is not a complete answer. AP2 needs adoption beyond Google's launch partners. UCP needs to become a true open standard rather than a Google-led one. Antigravity needs to ship reliable agentic experiences at billion-user scale, which it has not yet demonstrated. The Engadget reporter live-blogging the keynote put it bluntly: "And this is where I wonder if Google has lost the plot a bit. Who is asking for agentic coding in search?"

That is a fair question. But the underlying motion, from a search engine that sends you somewhere to a destination that does the work for you, is not in dispute. The only dispute is timing.

What to do this quarter

For Services and SaaS

1. Audit your dark AI traffic.

• Pull your GA4 direct-traffic trend, compare against branded search volume in Search Console. If direct is growing faster than branded, you are already getting AI referrals you cannot see.
• Add server-side logging or a third-party attribution tool to surface the real channel mix. Plan on this being 5 to 15% of total traffic once you can see it.

2. Restructure your highest-intent comparison pages for entity extraction.

• AI Overviews and AI Mode cite content that is unambiguous, structured, and depth-rich. One third of ChatGPT citations come from pages three folders deep in site architecture.
• Move from "rank for keyword" to "be the named source." FAQ schema, comparison tables, and primary research stand out. Thin top-of-funnel content does not.

3. Diversify beyond ChatGPT.

• Track brand mention share across ChatGPT, Gemini, Perplexity and Claude. Aim for at least 35% of AI mentions to come from non-ChatGPT sources by Q3.
• If 30% of your buyer base sits in Perplexity (senior leadership), being absent there is structurally more expensive than being absent on ChatGPT.

For ecommerce

1. Get your product data feed in order.

• If your Google Merchant Center feed has gaps, your Universal Cart eligibility will too. AP2 will only get more demanding.
• Audit attribute completeness, structured data, and image quality. The bar moves from "good enough for Google Shopping" to "good enough for an agent making a purchase decision on a user's behalf."

2. Pressure-test your brand visibility in AI-driven discovery.

• Ask ChatGPT, Gemini, Perplexity and Claude the queries your category lives on. If your brand is not in the answer, fix that before you fix anything else.
• Industry leaders like CeraVe (27% AI mention share in beauty) and Apple (54% in electronics) built that visibility deliberately.

3. Build for branded demand, not just intercepted demand.

• The CTR gap between branded and generic queries when AI Overviews appear is now 58 percentage points. Brand is the moat.
• Reinvest in upper-funnel brand work that the AI cannot synthesize away. PR, partnerships, owned audience, distinctive product. The performance-only stack is exposed.

The 6th Man take

Google's I/O 2026 keynote was an announcement and a confession at the same time. The announcement is that search is now an answer engine with a transaction layer attached. The confession is that the click, the unit Google built its $400 billion business on, is no longer the most important thing happening in search.

For Services and SaaS, that means the funnel starts upstream of the click. For ecommerce, it means the discovery layer has moved into ChatGPT, Gemini and Perplexity, and the checkout layer is becoming an open protocol. In both cases, brand recognition is the only thing that still does what SEO used to do.

The companies that act on this in 2026 will look prescient by 2028. The ones that do not will spend the next two years explaining why their traffic is down.

‍

Sources

1. A new era for AI Search.Liz Reid, Google Blog, May 19, 2026.
2. I/O 2026: Welcome to the agentic Gemini era.Sundar Pichai keynote, Google Blog, May 19, 2026.
3. Introducing the Universal Cart.Vidhya Srinivasan, Google Blog, May 19, 2026.
4. How AI Mode is changing the way people search in the U.S..Shivani Mohan, Google Blog, May 19, 2026.
5. Google Just Killed Search As We Know It.State of Brand, May 19, 2026.
6. Google gets 'intelligent' Search box redesign.9to5Google, May 19, 2026.
7. Google courts coders and consumers at I/O.Reuters via GBAF, May 19, 2026.
8. Google I/O 2026 Live.Tom's Guide, May 19, 2026.
9. Generative AI Statistics for 2026.Similarweb.
10. Gen AI Website Traffic Share Report.The Digital Bloom, February 2026.
11. Impact of AI Overviews on Publishers.Search Engine Journal.
12. Google AI Overviews publisher traffic decline analysis.ALM Corp, 2026.
13. State of AI Traffic 2026 Benchmark Report.Loamly.ai.
14. Google AI Overview Statistics: 2026 Trends.Heroic Rankings.
15. ChatGPT Commerce & Agentic Shopping Statistics 2026.Elogic Commerce.
16. Google I/O 2026 Live Blog.Engadget, May 19, 2026.

‍