AI Overviews are turning a large part of B2B organic traffic into a visibility and authority game before it becomes a click game. CMOs still need organic search, but they need to stop managing every query as if the desired outcome were a session. Ahrefs first found that AI Overviews correlated with a 34.5% lower average click-through rate for the top-ranking page in March 2025. When they reran the study using December 2025 data, the reduction had grown to 58% for position one content. That is large enough to change how a CMO reads the organic report. If a page used to earn 100 clicks from a top organic position, the current pattern suggests that 58 of those clicks may be resolved inside the search results before the buyer ever reaches the site. For a CMO, the business implication is bigger than “traffic is down.” Traffic can fall for many reasons. The useful interpretation is that Google is absorbing more of the informational journey. Definitions, explanations, “how does this work” queries, and early research questions are increasingly answered by the interface itself. The company can still influence the answer, but the value shows up as citation, mention, preference, and category association before it shows up as website traffic. That changes the budget conversation. Most B2B teams built their content programs around a simple organic path: rank, get the click, educate, retarget, convert later. That path still exists, especially for commercial and service pages. It is weaker for informational content. A buyer searching “what is revenue marketing operations” or “how to measure AI marketing ROI” may get a synthesized answer, a few citations, and enough context to keep moving without opening five tabs. If your page is part of the answer, it can still shape the buyer’s mental model. If your page is absent, your competitor or a neutral third-party source may define the category for you. The mistake is treating all organic traffic loss as the same kind of loss. This is the same lens we used in What to Review When AI Search Makes Funnel Signals Less Complete — read the same dashboard and you may misdiagnose the gap as a demand problem when it is really an evidence problem. A lost click on a pricing page hurts differently from a lost click on a definition page. A lost click on a “best agency for B2B SaaS demand generation” query has a different meaning than a lost click on “what is a marketing qualified lead.” The first may be close to vendor evaluation. The second may be an authority and education moment where the click was always a weak proxy for value.
The click loss is concentrated where education used to happen
Ahrefs’ original study focused on informational keywords because their earlier research found that 99.2% of keywords triggering AI Overviews were informational. A later Ahrefs trigger study put the figure at 99.9% of AI Overview keywords being informational, with much lower appearance across commercial, transactional, and navigational intent. That matters because many B2B content budgets are still arranged by content format, not intent. Blog, SEO, product marketing, and demand gen often sit in separate planning lanes, but the buyer does not experience those internal lines. The buyer asks a question, sees an AI answer, and continues with whatever source shaped the answer most clearly. When informational queries lose clicks, the job of that content changes. It still needs to be accurate, clear, differentiated, and visible. It also needs to be structured around the specific question the buyer is asking. The page has to make it easy for Google, ChatGPT, AI Mode, Perplexity, and third-party readers to understand the answer, the source of the claim, and the perspective behind it. That is a different standard from the old SEO playbook of writing one long “ultimate guide” and hoping it ranks for every adjacent term.
Commercial pages still need clicks
The citation era does not remove the need for conversion pages. Service pages, pricing pages, comparison pages, implementation pages, case studies, and buyer guides still need clicks because the buyer needs vendor-specific detail. AI can summarize a category. It cannot make a serious buying decision without the buyer eventually inspecting proof, process, scope, risk, pricing logic, and credibility. This is where many teams overcorrect. They see organic clicks falling, decide that the website matters less, and move budget into vague “AI visibility” activity. That usually creates another measurement problem. A B2B company still needs pages designed to convert demand when the buyer is ready to evaluate options. The practical move is to split the portfolio by query intent and desired business outcome. The portfolio map for clicks vs citations covers the operational version of this split in detail.
A content and query intent split for CMO budget decisions
| Query or page type | Buyer state | Primary outcome | What to measure | Budget posture |
|---|---|---|---|---|
| Definitions and “what is” explainers | Learning the category | Citation, mention, category association | AI citations, branded mentions, assisted search demand, inclusion in AI answers | Maintain and sharpen |
| Single-question explainers | Trying to solve a specific confusion | Being the clearest answer | Citation presence, answer accuracy, follow-on branded searches | Expand selectively |
| Category POV content | Building a point of view | Preference formation | Mentions in AI answers, third-party references, direct traffic to POV pages | Invest where SR has a strong stance |
| Research-backed claims | Looking for proof | Source authority | Citation rate, backlinks, earned mentions, sales usage | Invest if the claim supports positioning |
| Service pages | Evaluating a provider | Click and conversion | Organic clicks, form starts, booked calls, scroll depth, assisted pipeline | Protect and improve |
| Comparison pages | Shortlisting options | Trust and fit assessment | Conversion rate, sales conversations influenced, query coverage | Build where buyers compare alternatives |
| Pricing or packaging pages | Testing commercial fit | Qualification | Qualified conversions, sales acceptance, disqualification reasons | Keep in click-and-conversion lane |
| Generic ultimate guides | Broad research with weak intent | Usually unclear | Decay rate, AI citation absence, low assisted conversion | Rewrite, split, or retire |
| Low-judgment listicles | Outsourced opinion without real evaluation | Weak authority | Citation volatility, low trust, no sales usage | Kill or rebuild with actual judgment |
| Third-party profiles and reviews | Validating the brand externally | External proof | Mentions on cited domains, review quality, inclusion in category pages | Fund as authority infrastructure |
This table is useful because it stops the argument from becoming ideological. Organic, clicks, and blog traffic still matter, but each asset needs a clearer job than “get more sessions.” If the query is informational, the content should be judged by whether it can shape the answer. If the query is commercial, the page still needs to win the click and turn attention into a qualified next step. If the source is third-party, the job may be validation rather than traffic.
What should change in the operating model
The content review needs a new first question: what job is this query doing for the buyer? For informational queries, the team should inspect whether the answer is clear enough to be cited. That means tighter pages, cleaner definitions, stronger authorship, named frameworks, original examples, and source-backed claims. It also means distribution beyond the company blog, because AI systems draw from a wider set of sources than a company’s owned domain. For commercial queries, the team should protect the conversion path. A service page about an AI visibility review should not be turned into a 4,000-word encyclopedia entry just because informational pages are losing clicks. It should help a serious buyer understand the work, the operating model, the fit, the proof, and the next decision. For third-party validation, the team should build a source map. Which review sites, industry publications, comparison pages, podcasts, YouTube videos, partner pages, and analyst-style resources show up in AI answers around the category? If those sources influence the model’s answer, they deserve a place in the budget discussion.
The CMO decision
The old content budget often rewarded volume because volume was easier to manage than judgment: more pages, more keywords, more traffic, more dashboards. AI Overviews make that model less honest. A page can lose clicks and become more strategically important if it is cited in the answer that defines the buyer’s problem. Another page can gain traffic and still create no business value if it attracts broad curiosity without buying intent. The CMO playbook for the citation era is a reallocation exercise:
- Keep commercial, service, comparison, and pricing pages in the click-and-conversion lane.
- Move informational content into a citation-and-authority lane.
- Fund third-party validation as part of search strategy, not as a separate PR nice-to-have.
- Review content by query intent before judging performance by traffic.
- Connect every visibility metric to a business interpretation, such as buyer education, preference, qualification, or pipeline influence.
The traffic graph will keep getting noisier. The operating decision has to get clearer.
Next read: Why Schema Markup Will Not Save Your AI Citations — the next piece in this series. If you want a structured way to inspect where AI search is already absorbing your funnel, see our methodology.
— Fernando González Aguirre, Founder, Structured Rebellion
