How to Measure and Defend Visibility in AI Search: A 4-Layer Framework and a 2-Week Test for B2B Marketers
The B2B SEO market has just entered a stage where the classic dashboard no longer tells the truth about a brand’s real visibility. As of August 31, 2026, Google rolled out Search Console reports for generative features globally, along with a control toggle for presence in AI Search—but at the same time, it did not provide the most important metric everyone was used to: clicks. This is not a cosmetic UI change. It’s a measurement-system shift that forces a new decision-making logic.
What exactly changed on Google’s side
Google launched two critical operational components:
Generative AI performance report in Search Console (for Search): shows, among other things, impressions, pages, countries, devices, and dates for AI Overviews and AI Mode.
Search generative AI control: lets you enable or disable your site’s presence in generative features (AI Overviews, AI Mode, gen-AI features in Discover).
The most important implications for B2B marketing teams:
There is no separate click metric in this report—we are talking about exposure measurement, not traffic.
Excluding a site means no traffic and no impressions from these features.
Setting changes usually take effect after a few days (typically 1–2 days, sometimes longer due to cache and propagation).
The control works hierarchically (parent/child property inheritance), so it’s easy to accidentally test the wrong thing.
A clickless measurement framework: 4 layers that work
If clicks disappear from your attribution model, you need leading indicators. Below is a practical framework you can implement immediately.
Layer 1: Presence (are we visible at all)
Use the AI Performance report for continuous monitoring:
Total impressions (weekly trend)
Share of pages with AI exposure
Breakdown by countries and devices
Critical methodological note:
The chart and table may differ due to a different aggregation method.
In some cases, multiple results from the same site within an AI feature are counted as one impression at the aggregate level.
This protects you from a classic error: artificially inflating a “visibility win.”
Layer 2: Query clusters (which intents are we competing for)
Don’t measure single keywords. Group queries by intent and funnel stage:
Informational “how/what/why”
Comparative “X vs Y”
Problem-solving “how to solve…”
Brand vs non-brand
Why this matters:
Ahrefs research shows a “great decoupling”—impressions rising while clicks fall.
AI Overview presence correlates with CTR declines (Ahrefs reported drops of around 34.5% in a large-sample analysis).
An independent study (arXiv, panel of 900 US adults) indicates that clicks on sources cited in AI Overviews are very rare (about 1% of visits on AI Overviews).
Conclusion: a cluster that “looks worse” in clicks may still be critical for influencing purchase decisions.
Layer 3: Page types (which page types actually carry visibility)
Instead of reporting on the domain as a whole, split URLs into types:
Educational TOFU pages
Comparative MOFU pages
Use-case / solution pages
Documentation / knowledge base
Commercial and product pages
Then monitor for each type:
AI impression dynamics
Country/device share
Exposure stability over time
This is the moment when the content team gets a clear signal: “we create more of what earns presence in AI answers,” not “we create more of whatever used to drive traffic.”
Layer 4: Business proxies (how to connect this to pipeline)
Because the AI report has no clicks, and clicks are increasingly weak at describing impact anyway, connect exposure data with business proxies:
Branded direct traffic by countries with growing AI impressions
Growth in branded queries
User return rate
SQLs/MQLs from thematic segments corresponding to query clusters
This is where you defend budget: you show not “fewer clicks,” but “greater share in brand-discovery moments.”
2-week opt-out test: how to quantify the cost of invisibility
This is the most important operational experiment right now. Its goal: quantify what you actually lose when you disable participation in generative features.
Recommended setup: quasi A/B on child properties
Choose two as-similar-as-possible segments:
e.g., subfolders or subdomains with similar seasonality and intent.
Define a baseline period:
minimum 14 days before the change.
In treatment, set:
Exclude my site's links and content from Search generative AI features.
Leave control unchanged.
Wait for propagation:
usually 1–2 days.
Measure for 14 full days:
difference in AI impressions,
impact on branded/direct/pipeline proxies.
What to compare (and how not to fool yourself)
Don’t look only at total organic traffic.
Compare delta treatment vs delta control (difference-in-differences).
Segment by:
country,
device,
page type,
query cluster.
If treatment loses AI impressions while control maintains trend, you have hard evidence of opt-out cost. If branded signals also drop in those same segments, you can quantify a business effect, not just an “SEO effect.”
How to report this to leadership without a defensive narrative
In executive communication, shift the conversation from “traffic down” to market visibility economics:
Old metric: clicks.
New reality: presence in the answer before the click happens.
Management decision: where it pays to defend AI visibility and where to reduce investment.
The strongest message for leadership:
Google provided an official AI exposure report and an official opt-out toggle.
The market is observing a structural divergence between impressions and clicks.
A 2-week experiment gives a quantifiable answer for your specific brand, instead of speculation based on case studies from other industries.
Over the next few quarters, the winners won’t be the teams that complain loudest about zero-click. They’ll be the teams that build their own AI Search impact measurement system fastest—and make decisions based on data, not nostalgia for historical CTR.
Sources
Google Search Console AI performance reports and Search generative AI control rolling out globally
Introducing Search Generative AI performance reports in Search Console
How to Track AI Overviews: Mentions, Citations, Click Loss, and the Traffic Google Won’t Show You
The Great Decoupling (or Why Your Clicks Are Down and Impressions Up)
Investigating Click Behaviors On Google Search Result Pages That Produce an AI Overview