Sundar Pichai called the redesign of Google's search interface at I/O 2026 the biggest change to Search in more than 25 years, and for once the hyperbole was roughly earned. That's the headline event of the year, but it's not the only thing that moved. I want to walk through what actually changed in the last twelve months, honestly, with sources attached, rather than repeating whichever single stat happens to serve a specific vendor's pitch.
How Much Has B2B Buyer Behavior Actually Shifted This Year?
Substantially, and the numbers keep climbing rather than plateauing. 94% of B2B buyers used generative AI or conversational search somewhere in their 2025 purchase process, up from 89% the year before, according to Forrester. Buyers named generative AI their single most meaningful research source, roughly twice as influential as any other channel measured. A separate benchmark specific to B2B technology buyers found 58% now use AI-powered search during initial vendor research, up from just 17% in 2023, a 3.4x increase in under three years.
Gartner's own late-2025 survey of 645 B2B buyers found a more nuanced picture worth holding onto: 45% used GenAI primarily to gather vendor information, but 69% still turned to a human sales rep to validate what the AI told them, drawing on an average of seven information sources per purchase. The honest read here isn't that AI replaced the rep. It's that AI now frames the conversation before the rep ever joins it. If your brand was framed out of that early research, the rep is closing a deal your company was never actually considered for.
What Was the Single Biggest Structural Change This Year?
Google's own AI Optimization Guide, published inside Search Central's official documentation on May 15, 2026. This was the first time Google committed to paper, in a permanent reference document, exactly what it does and doesn't reward for AI Overviews and AI Mode visibility. The core claim, that AEO and GEO are "still SEO" from Google's own perspective, reshaped a meaningful share of the industry conversation this year, and its mythbusting section specifically dismissed llms.txt, content chunking, and AI-specific rewriting as unnecessary for Google's own systems.
That guidance sits alongside a separate, less publicized but genuinely relevant data point: 97% of llms.txt files tracked in one large-scale analysis received zero traffic as of May 2026. The two facts corroborate each other. Whatever value llms.txt has, and it may have some for AI coding agents specifically, it is not currently the visibility lever much of the industry treated it as through 2025.
What Changed at the Platform Level?
Google AI Mode crossed 1 billion monthly users this year, with the interface itself redesigned around what Google has confirmed is fundamentally different query behavior: AI Mode searches now run roughly three times longer than traditional queries, and follow-up questions within a single session are up 40% month over month. AI Overviews now appear on roughly a fifth to a quarter of all tracked keywords, depending on the measurement source, up sharply from the prior year.
Perplexity had the most dramatic repositioning of any single platform this year. It stopped describing itself primarily as "a better search engine" and repositioned explicitly as an agentic AI platform with search as one core capability rather than the whole product. Annualized revenue grew 335% year over year from roughly 148 million dollars in mid-2025, crossing an estimated 450 to 500 million dollars by the following spring, driven substantially by its new agentic product line. A Snapchat integration extended Perplexity's answer engine to a user base approaching one billion, well beyond its own standalone app.
ChatGPT, meanwhile, remains the largest single AI chatbot by user base, but the data on how much of a brand's actual visibility runs through it specifically is more complicated than raw user counts suggest, a point worth returning to below.
Does Traditional Organic Ranking Still Predict AI Citation?
Less than most SEO teams assume, and the gap has if anything widened this year. One large-scale study found 76.95% of AI-cited URLs were not present in the organic top ten for the same query. A separate analysis found AI Overview presence cuts the position-one organic clickthrough rate by roughly 58%, meaning even a page that does rank first is capturing meaningfully less of the traffic it used to when an AI Overview sits above it.
Zero-click behavior overall has climbed from around 50% in 2019 into the mid-60% range by 2026. For queries specifically triggering an AI Overview, zero-click behavior runs even higher, and inside AI Mode and standalone chat assistants, a majority of sessions now end with no website visit at all. This is the uncomfortable core fact underneath everything else in this piece: a meaningful and growing share of buyer research now happens without your website ever being visited, whether or not your brand was the one cited.
What's the Actual Conversion Story for the Traffic That Does Arrive?
Better than organic, consistently, across every source measuring it, even though the exact multiple varies. Depending on the study, AI-referred traffic converts somewhere between roughly 4 and 5 times the rate of standard organic search traffic, with individual reports ranging from 11.4% to 14.2% for AI-referred visits versus 2.8% to 5.3% for organic. The specific number moves depending on methodology and vertical. The direction is completely consistent across every independent source: a buyer arriving via an AI citation has typically already been pre-qualified by the synthesized answer they read, and lands closer to a decision than a buyer clicking a traditional blue link.
Why Are Most B2B Marketing Teams Still Behind on This?
Measurement, mostly. Only 22% of marketers currently track AI visibility at all, and 64% say they don't know how to measure AI search performance even if they wanted to start. Google's own free Search Console reporting will now confirm when a page appears inside an AI Overview, but it still won't tell you whether anyone clicked, or what question triggered the appearance, which leaves teams able to confirm visibility without being able to confirm it mattered.
This measurement gap compounds a second, more structural mistake that became clearly visible this year: gating high-authority content behind an MQL form. Content locked behind a form produces zero AI citation share, because a system retrieving and citing sources can't cite what it can't access. By the middle of this year, most leading B2B SaaS companies had un-gated their best-performing content specifically for this reason, and the companies still gating it are visibly losing citation share to competitors who made the switch.
What Do Reddit and Review Platforms Have to Do With Any of This?
More than most B2B marketing budgets currently reflect. Across independent research this year, third-party corroboration, Reddit discussion, G2 and Capterra reviews, industry press, consistently outweighs a brand's own website content in what actually earns an AI citation. Most B2B companies have neglected these specific platforms for years, treating them as a minor rating-page afterthought rather than a primary visibility channel. Building out a genuinely complete, current G2 and Capterra presence is, by a wide margin, one of the highest-return, least-glamorous fixes available to almost any B2B team reading this.
Is There Any Real Academic Consensus on What Actually Works?
Less than the industry's confident tone usually suggests, and it's worth saying so plainly rather than overselling certainty that doesn't exist. A careful academic review published in July 2026, surveying 45 separate studies, concluded that no single reviewed technique could be shown to conclusively and reliably move AI citation outcomes across contexts. That's a genuinely useful caveat to hold onto, not a reason to do nothing. It means the tactics with the strongest, most repeatedly observed directional support, un-gating high-value content, building genuine third-party corroboration, structuring content around direct, specific answers, deserve real investment, while anyone promising a guaranteed, deterministic formula is overstating what the current evidence actually supports.
What Should a B2B Team Actually Prioritize Going Into Next Year?
Start measuring AI visibility at all, even manually, since the 22% of teams already doing this have a genuine head start over the 78% still flying blind. Un-gate the content that's actually good enough to earn a citation, since gated authority content is invisible to every AI system by definition. Build out third-party presence on the review platforms that consistently outweigh owned content in citation research. And hold realistic expectations about certainty in this field specifically, since the most rigorous academic review available this year found no technique that works reliably across every context. Direction matters more than any single tactic promising a guaranteed result.
Frequently Asked Questions
What was the single biggest change to AI search in the last 12 months?
Google's official AI Optimization Guide, published May 15, 2026, which for the first time stated directly that AEO and GEO are "still SEO" from Google's own perspective, and explicitly dismissed several popular tactics, including llms.txt and content chunking, as unnecessary for its own AI Overview and AI Mode systems. This reshaped a meaningful share of industry strategy this year, particularly once independent data confirmed 97% of llms.txt files were receiving zero traffic.
Does ranking well in Google still predict AI citation in 2026?
Less reliably than in prior years. One large-scale study found 76.95% of AI-cited URLs were not present in the organic top ten for the same query, and AI Overview presence has been shown to cut position-one organic clickthrough by roughly 58%. Traditional ranking remains relevant but is no longer sufficient on its own to predict or guarantee AI citation.
How much has B2B buyer use of AI search actually grown this year?
Significantly. 94% of B2B buyers used generative AI somewhere in their 2025 purchase process, up from 89% the year prior, according to Forrester. A separate B2B technology-specific benchmark found 58% now use AI-powered search during initial vendor research, up from 17% in 2023.
Why do most B2B marketing teams still lag on AI search visibility?
Primarily measurement gaps and gated content. Only 22% of marketers currently track AI visibility, and 64% say they don't know how to measure it. Compounding this, content gated behind lead-capture forms produces zero AI citation share, since AI systems can't cite what they can't access, and many B2B companies have only begun un-gating high-authority content this year specifically because of this realization.
Is there solid academic consensus on which AI search tactics actually work?
Not yet, and it's worth being honest about that rather than overselling certainty. A July 2026 academic review of 45 studies found no single technique could be shown to reliably move AI citation outcomes across every context. The tactics with the strongest repeated directional support, un-gating valuable content, building genuine third-party review presence, and structuring content around direct answers, are worth real investment, but no current research supports a guaranteed, deterministic formula.
References
79 Development
The State of AI Search 2026: Data & Insights, Google I/O 2026 redesign details and Perplexity's agentic repositioning and revenue growth: https://79dev.com/state-of-ai-search-2026/
Andrii Byzov
The State of AI Search for B2B SaaS in 2026, Forrester and Gartner buyer behavior data and zero-click search trend analysis: https://blog.andrewbyzov.com/posts/state-of-ai-search-for-b2b-saas-2026/
OrganikPI
The State of AI Search in 2026: 150+ Statistics, 76.95% non-top-10 citation data and llms.txt zero-traffic finding: https://organikpi.com/blog/geo-ai-search/state-of-ai-search/
GEO Compass
The State of AI Search, Mid-2026: What Has Actually Changed for Publishers, content-gating citation impact and vendor measurement landscape: https://guptadeepak.com/geo-compass/guides/state-of-ai-search-2026/
Anurag Pareek
AI Search Statistics for B2B in 2026: What Actually Changed, review platform citation weighting and AI visibility measurement gap data: https://www.anuragpareek.com/blog/ai-seo-statistics/
Kalungi
The State of Organic and AI Search for B2B SaaS, Q3 2026 Update, academic review of 45 studies on AI citation technique reliability: https://www.kalungi.com/blog/the-state-of-organic-and-ai-search-for-b2b-saas-august-september-2026-update
