Search visibility used to be easier to measure. If rankings improved, impressions increased, clicks followed, and organic traffic became the main proof of progress. AI search complicates that model. New experimental research suggests that AI-generated search experiences can reduce external clicks even when users continue searching and consuming information. For brands, this means the question is no longer only “Did we get the click?” It is also “Were we included in the answer?”
Seonoob is an SEO and AI visibility platform built to help teams monitor both dimensions, tracking where brands appear in traditional search results and in AI-generated answers across the platforms where their audiences now spend time.
What the New AI Search Experiment Found
A preregistered field experiment published on arXiv in August 2026 provides some of the strongest experimental evidence yet that AI search features reshape how users engage with the broader web. The paper, titled “AI in Search Reduces Publisher Referrals Without Improving User Experience: Experimental Evidence”, was authored by researchers at the University of Pennsylvania and Northeastern University and submitted as a version-one preprint on 18 August 2026. It has not yet completed peer review, and results should be treated as early but important signal rather than settled findings.
The Study Design
The study was a preregistered randomized controlled trial involving 1,100 Google users. Participants were recruited between March 17 and 19, 2026, from the research platform Prolific and from Northeastern University. All participants were US-based, 18 or older, used Google Chrome as their primary browser, and used Google Search as their main search engine.
After a three-day baseline period, participants were randomly assigned to one of three conditions for seven days:
- No AI Search, which hid AI Overviews and redirected away from AI Mode
- Current Search, which made no modifications to Google Search
- AI Mode Search, which redirected all searches to AI Mode
The researchers compared behavioral outcomes including click-through rates, search sessions, and clicks to specific domains, alongside survey-based measures of trust, usefulness, satisfaction, and agency. This experimental design allows the researchers to isolate the effect of AI features from other variables, which is a meaningful methodological strength compared to observational studies.
The Click-Through Finding
The headline result is direct. Assignment to AI Mode reduced external click-through rate by 18.8 percentage points, and the finding was statistically significant at p less than 0.001. In the other direction, removing AI features entirely increased external click-through by 8.8 percentage points.
Data callout: In the preregistered experiment, AI Mode reduced external click-through by 18.8 percentage points. Removing AI features increased external click-through by 8.8 percentage points.
This is not a small effect. It means that users who were routed into Google’s conversational AI search experience clicked through to external websites far less often than users in the standard search experience, without any gain in satisfaction to explain the difference.
For publishers, content-led brands, review sites, education providers, legal content creators, SaaS blogs, and ecommerce discovery content, these numbers represent a structural challenge to traffic-based reporting. The users are still searching. They are spending more time within the AI interface, not less. But referral traffic to the broader web is declining as a result.
The User Experience Finding
An important nuance in the paper is that AI Mode did not simply reduce clicks while making users happier with the experience. The researchers found that AI Mode reduced perceived trust in information on Google, reduced perceived usefulness, satisfaction, and agency, and also reduced perceived personalization and relevance of responses. These are all statistically significant decreases compared with standard Google Search.
Users in the AI Mode condition also showed significantly higher rates of switching to competitor search engines, with an 11.2 percentage point increase in the fraction of users turning to Bing, DuckDuckGo, or Yahoo. That finding matters for how AI Mode’s future is likely to evolve.
Marketers should not assume that fewer clicks always mean users are happier or better served. The experimental evidence points in the opposite direction: users are spending more time inside the AI interface while rating the experience less favourably than standard search.
The Publisher and Source Impact
Beyond overall click-through, the study measured the fraction of users clicking through to specific domain categories. AI Mode significantly reduced clicks to news sites by 12.5 percentage points, to Reddit by 21.2 percentage points, and to Wikipedia by 9.9 percentage points. The paper concludes that “referral traffic to the broader web declines substantially,” and that these effects generalise well beyond any single website type.
This is relevant for every business category that has relied on informational search to drive discovery. The impact is not narrowly confined to news publishers. Content-led brands of every kind face a version of the same pattern.
Important Limitations
The study has limitations that should be kept in mind before drawing strong conclusions. Compliance in the No AI Search condition fell to roughly 50% overall due to a Google interface change during the study period, so the researchers used local average treatment effect estimates for that condition rather than simple intent-to-treat comparisons. The sample skewed younger, more educated, and more left-leaning than the general US population. The experiment covered seven days of treatment, so longer-term adaptation effects are unknown. And baseline AI Mode usage was very low at 0.6% of searches, meaning the AI Mode condition reflected forced full adoption rather than the gradual opt-in experience most users would have in practice. The researchers also note that AI Mode had not yet monetised through advertising at the time of the study, which may affect future results.
This research should guide measurement strategy and planning. It should not be used to make exaggerated claims about the end of organic search.
Why This Matters for SEO Reporting
SEO reporting has traditionally connected visibility to traffic. Rankings improve, clicks increase, sessions climb, and conversions can be tracked. AI search breaks part of that connection because visibility may now happen inside the answer itself, before the user ever decides to click.
A brand can be mentioned, cited, summarised, or compared in an AI answer without producing a session in Google Analytics. Referral traffic may underreport the real influence of AI search, while organic sessions alone can make AI visibility look weaker than it actually is. At the same time, citations without business outcomes can be overvalued if they never translate into demand.
The solution is a layered reporting model that separates visibility from traffic, then connects both to downstream business outcomes.
Old SEO model: Ranking improves → clicks increase → sessions increase → conversions are measured.
AI search model: Brand appears in AI answer → user may not click → brand recall or trust changes → later branded search, direct visit, assisted conversion, or offline decision.
In AI search, the path from visibility to revenue is longer and less direct. Measuring only the final click misses most of what happened in between.
AI Visibility and Website Traffic Can Diverge
A Brand Can Be Visible Without Receiving the Click
AI answers can include a brand mention or citation without producing a session in any analytics platform. The user may have received enough information inside the AI interface and moved on. The brand still gained awareness. The analytics dashboard may show nothing.
Consider a user asking: “What are the best SEO tools for tracking AI visibility?” If an AI answer mentions Seonoob by name but the user does not click immediately, Seonoob still gained a visibility impression. That impression may later influence a branded search, a direct visit, a trial signup, or a word-of-mouth recommendation. None of those downstream effects would show as an AI referral in most current analytics setups.
A Cited Page May Not Be the Converting Page
AI systems may cite blog posts, research pages, comparison pages, help articles, or product pages depending on what is most relevant to each part of an answer. A cited informational page often influences discovery, while conversion happens later through the homepage, pricing page, or a direct search. Reporting that focuses only on cited-page performance misses this downstream chain.
AI Mentions May Influence Demand Before Traffic Appears
Users who encounter a brand in an AI answer may search for it later, ask follow-up prompts in the same AI session, compare competitors inside the same interface, or click a different source that discusses the brand. AI visibility should be measured over time as a discovery signal, not only as an immediate referral.
Who Is Most Exposed to AI Search Click Loss?
Content-Led Businesses
Businesses whose marketing depends heavily on informational content are most directly affected. Educational content often answers top-of-funnel questions. AI Overviews and AI Mode are most likely to satisfy informational intent inside the search interface, meaning fewer users need to click through to read the full article. Training providers, legal firms, SaaS blogs, health information sites, and publishers of all kinds fall into this category.
Being cited as a supporting source in an AI answer still has value for these businesses. It shapes brand perception and recommendation patterns even when no click follows.
Professional Services
Legal, finance, consulting, and healthcare-adjacent businesses often rely on informational search to warm up audiences before a conversion. AI search may now answer basic questions before users contact a firm. For a firm advising on complex matters like offshore structures, cross-border disputes, or sector-specific regulation, being included as a trusted source in AI answers can still shape the consideration stage, even without the direct click. Trust, source inclusion, author credibility, and brand mentions become more important as the click itself becomes less reliable as a measurement tool.
Ecommerce Research Queries
Ecommerce discovery often begins with comparison searches. AI answers may summarise product types, pros and cons, and brand comparisons without requiring a click to any individual retailer. Brands in categories like engagement rings, electronics, apparel, and home goods need to monitor whether they are mentioned in comparison-style and recommendation-style prompts. Being excluded from these answers is a real business problem that will not show up in standard traffic data.
SEO and SaaS Tools
SaaS buyers increasingly ask AI tools to compare software options before visiting any product page. A tool can gain or lose pipeline before the user ever reaches the website. This makes LLM rank tracking particularly important for SaaS marketers, because the conversations that shape purchase decisions are happening in AI interfaces, not only on Google’s blue-link results page.
Why Referral Traffic Is No Longer Enough
Standard analytics platforms measure visits that happen after a click. AI search can influence users before the click, during a session that produces no referral. Some AI platforms do not pass clean referral data in any case. And users who encounter a brand in an AI answer often convert through branded search, direct visits, email, or paid search after the initial exposure.
AI visibility should be treated like a discovery and influence layer, not solely as a referral channel. That means separating visibility metrics from traffic metrics in reporting.
| Traditional Metric | What It Shows | What It Misses in AI Search |
|---|---|---|
| Organic clicks | Visits from search results | Brand mentions without clicks |
| Organic sessions | Website traffic from search | AI answer exposure inside search |
| Referral traffic | Visits from external platforms | No-click influence |
| Keyword rankings | Position in search results | AI prompt-level visibility |
| Conversions | Final action taken | Assisted discovery and brand recall |
| Impressions | Search result exposure | Whether AI mentioned or cited the brand |
The New AI Search Reporting Model
Layer 1: Prompt Visibility
Track whether the brand appears for a fixed set of prompts across relevant intent categories:
- Informational prompts
- Recommendation prompts
- Comparison prompts
- Alternatives prompts
- Local prompts
- Pricing prompts
- Problem-solving prompts
- Industry-specific prompts
For an SEO platform, example prompts might include: “What are the best AI visibility tools for SEO?”, “Which tools track LLM mentions?”, “Best keyword rank checker for small businesses”, “How can I measure brand visibility in ChatGPT?”, and “Best tools for tracking AI search mentions.”
Layer 2: Citation Share
Track whether the brand’s own site is cited, whether third-party pages mentioning the brand are cited, what competitors are being cited, and what source types appear most often (blog, homepage, review page, video, directory, community).
Citation share: The percentage of tracked AI responses where the brand, its website, or a supporting third-party source is cited.
Layer 3: Brand Mention Share
A brand may be named in an AI answer without receiving a citation. Track how often the brand appears relative to competitors, whether the mention is positive, neutral, or negative, and where in the response the brand appears.
Mention share: The percentage of tracked AI responses where the brand is named.
Layer 4: Sentiment and Context
Being mentioned is not always useful. AI may describe a brand as expensive, limited, outdated, or suitable only for particular use cases. Track how the brand is framed and whether those descriptions are accurate. Correct inaccurate positioning through better product pages, clearer feature descriptions, and stronger third-party coverage.
AI sentiment score: A qualitative or categorised measure of how AI describes the brand across a fixed prompt set.
Layer 5: Downstream Outcomes
Connect AI visibility to business outcomes by tracking branded searches, direct visits, assisted conversions, trial signups, demo requests, returning users, newsletter signups, lead quality, and sales conversations where AI-generated discovery is mentioned. Customer surveys asking “Where did you first hear about us?” can also surface AI as a discovery channel that standard attribution would miss.
AI search should not be judged by referral traffic alone. Brand demand, source inclusion, assisted conversions, and mention quality are all part of the measurement model.
How to Run a Visibility-to-Outcome Test
Step 1: Build a Fixed Prompt Set
Create between 25 and 100 prompts across key intent categories. Include category prompts, problem prompts, product prompts, brand comparison prompts, competitor prompts, best-of prompts, local prompts, and pricing prompts. Consistency matters here. Using the same prompts over time is the only way to detect real changes in visibility rather than natural variation.
Step 2: Track AI Citation and Mention Share
For each prompt, record the AI platform, date, prompt text, whether the brand was mentioned or not, which competitors were mentioned, whether the brand website was cited, which third-party sources were cited, mention position, sentiment, and accuracy.
Step 3: Monitor Downstream Demand
Track for four to six weeks: branded search impressions and clicks, direct traffic volume, organic assisted conversions, trial signups, demo requests, returning users, conversion paths, lead source notes, and customer survey responses.
Step 4: Compare Cited and Uncited Landing Pages
Group landing pages into four categories: cited by AI, mentioned but not cited, not cited or mentioned, and competitor cited instead. Compare organic sessions, branded search movement, assisted conversions, engagement, conversion rate, content freshness, and third-party mention volume across each group.
Step 5: Report Visibility Separately From Traffic
Use Seonoob’s AI visibility reporting workflow to create a dashboard with separate sections for AI visibility, AI citations, brand sentiment, competitor presence, organic traffic, branded demand, assisted conversions, and recommended actions. Keeping these separate prevents AI visibility from being obscured by traffic metrics and vice versa.
What This Means for Content Strategy
Create Content That Can Be Cited, Not Just Ranked
AI systems need clear, useful, and verifiable information. Content more likely to earn citations includes original research, comparison guides, definition pages, data-backed explainers, product use case pages, expert-led articles, FAQ sections, methodology pages, case studies, and first-party statistics.
The goal is not to produce content that tricks AI into mentioning the brand. It is to build pages that are so directly useful and clearly structured that AI systems have a genuine reason to include them.
Strengthen Entity and Brand Signals
Make the brand’s positioning clear across the site. Explain who the product is for, keep feature pages updated, maintain consistent naming, earn third-party mentions, build author and company credibility, and make pricing and product capabilities easy to understand. These signals collectively help AI systems build an accurate picture of what the brand does and for whom.
Update High-Value Informational Content
AI search may summarise informational content without always driving a click. That does not make informational content useless. The goal is to make it strong enough to be cited, remembered, and used as a supporting source. Update high-priority pages with original insights, examples, visuals, FAQ sections, and source references. Keep statistics and product details current.
Build Content for Follow-Up Questions
AI search encourages conversational follow-ups. For each main topic, build supporting content around: how it works, pros and cons, best tools, common mistakes, examples, pricing, comparisons, alternatives, and industry use cases. This breadth supports the query fan-out process and increases the likelihood of being cited across a wider range of related prompts.
What This Means for Different Business Types
For Publishers
Expect some informational click loss as AI answers absorb the top layer of search demand. Focus on loyal audiences, newsletters, direct visits, and building brand authority that generates return visits independent of search. Track which articles are being cited as sources in AI search and use that as evidence of content quality even when clicks are not coming from those citations.
For SaaS Brands
Track comparison prompts and recommendation prompts as priority metrics. Improve product pages, use case documentation, pricing clarity, and comparison content. Monitor whether AI tools describe the product accurately and completely. Measure branded search lift as a proxy for AI-influenced discovery rather than waiting for direct AI referrals that may never arrive.
Use a dedicated keyword research tool to build out a full content map across the informational, comparison, and decision-stage topics where AI systems are most likely to generate answers relevant to the product category.
For Agencies
Update SEO reporting templates to include AI visibility as a separate measurement layer. Add prompt tracking to monthly reporting cycles. Compare client and competitor mention share across a fixed prompt set. Educate clients that lower informational click counts may not always indicate weaker visibility, and that AI influence often precedes rather than accompanies the click.
For Ecommerce
Monitor product category prompts. Track brand inclusion in buying guides and AI-generated comparison responses. Strengthen review visibility, product education content, and third-party mentions. Use AI visibility data to inform content planning and digital PR targeting.
For Local and Professional Services
Track local recommendation prompts. Improve service pages, author bios, FAQ sections, reviews, and location signals. Monitor inaccurate AI answers, which occur more often than most brands realise, and build trust signals beyond the website through directories, professional associations, case studies, and press coverage.
Recommended AI Visibility Metrics to Add to SEO Reports
| Metric | What It Measures | Why It Matters |
|---|---|---|
| Prompt visibility | Whether the brand appears for tracked prompts | Measures AI answer inclusion |
| Citation share | How often the brand or its pages are cited | Measures source inclusion |
| Mention share | How often the brand is named | Measures brand presence |
| Competitor share | Which competitors appear instead | Shows visibility gaps |
| Sentiment | How the brand is described | Identifies positioning issues |
| Accuracy | Whether AI descriptions are correct | Prevents misinformation |
| Source type | Which pages or domains are cited | Guides content and PR strategy |
| Branded search demand | Follow-up interest after AI exposure | Connects visibility to demand |
| Assisted conversions | Later conversions influenced by discovery | Connects AI visibility to revenue |
| Referral traffic | Actual visits from AI platforms | Useful, but incomplete alone |
The key is to separate visibility from traffic, then connect both to business outcomes. AI visibility is an upstream signal. Traffic and conversions are downstream signals. A complete report shows both.
Common Mistakes Marketers Should Avoid
Mistake 1: Judging AI SEO Only by Referral Traffic
Referral traffic misses every brand impression that happened inside an AI answer without producing a click. The experimental evidence suggests this is now a substantial share of AI-influenced discovery.
Mistake 2: Tracking One Prompt Once
AI answers vary across sessions, platforms, time, and query phrasing. A single check produces a data point, not a trend. Track a fixed prompt set on a consistent schedule to detect real changes.
Mistake 3: Ignoring Competitor Mentions
Competitor visibility can be rising in AI answers even while a brand’s own traffic looks stable. Monitoring what AI says about competitors reveals market-share dynamics that standard analytics cannot show.
Mistake 4: Treating Citations as Conversions
A citation is a visibility signal. It needs to be connected to downstream business metrics before it has commercial meaning. Being cited in an AI answer is a starting point, not a result.
Mistake 5: Creating Generic AI Content
Generic content is easy for AI to summarise and move on from, with no attribution, no citation, and no brand mention. Original research, specific examples, clear expertise, and verifiable claims are more likely to create defensible visibility in AI-generated answers.
How Seonoob Fits Into the Clickless Search Era
Seonoob is built for this shift. As AI search changes how people discover brands, marketers need tools that show more than keyword rankings and traffic charts. The platform helps teams track where brands appear across AI-generated answers, monitor how competitors are being represented, and identify the prompts where brand visibility is missing.
LLM rank tracking lets users monitor brand mentions and citations across AI platforms at the prompt level, giving marketers the data they need to understand AI visibility as a distinct measurement layer alongside traditional search performance.
As the new field experiment makes clear, the gap between visibility and traffic is widening. A brand can be present and influential in AI answers without those answers producing immediate referral sessions. Connecting those upstream visibility signals to downstream demand, branded search growth, and assisted conversions is the reporting challenge that defines modern SEO.
Final Takeaway: The Future of SEO Reporting Is Visibility Plus Outcomes
AI search does not remove the need for SEO. It changes what needs to be measured. Rankings, clicks, and sessions still matter, but they no longer tell the full story on their own.
The experimental evidence from Wang, Gleason, Bart, Wilson, and Metaxa (2026) is preliminary, but the direction it points is consistent with what many marketers are already observing in their own data. AI search can reduce external click-through while keeping users engaged inside the search experience. That pattern has real consequences for every business that has treated organic sessions as the primary measure of search success.
In the clickless search era, brands need to know whether they are included in AI answers, cited as a source, described accurately, and chosen over competitors. The next generation of SEO reporting will connect visibility to outcomes, not treat traffic as the only proof of success.
Frequently Asked Questions
Does AI search reduce website traffic?
New experimental research suggests that AI search can reduce external click-through. In one preregistered field experiment, AI Mode reduced external click-through by 18.8 percentage points, while removing AI features increased click-through by 8.8 percentage points. This is preliminary evidence from a preprint and should be interpreted carefully.
Does this mean AI visibility is more important than SEO traffic?
No. AI visibility and SEO traffic should be measured together as complementary signals. AI visibility shows whether a brand appears in AI answers, while traffic shows whether users clicked through. Both are important, but they measure different parts of the discovery journey.
What should brands track in AI search?
Brands should track prompt visibility, citation share, brand mention share, competitor presence, sentiment, accuracy, source inclusion, branded search demand, and assisted conversions.
Can a brand benefit from AI search without getting a click?
Yes. A user may encounter a brand in an AI answer, remember it, search for it later, visit directly, or convert through another channel. That is why AI visibility should not be judged only by referral traffic.
How can Seonoob help with AI visibility tracking?
Seonoob helps marketers track how their brand appears across AI-generated answers, monitor LLM mentions, compare competitor visibility, and connect AI visibility insights with broader SEO reporting. It is designed to support both traditional and AI-era measurement from one platform.
