Why AI overviews are killing your B2B SaaS organic traffic

AI overviews are killing your B2B SaaS organic traffic by answering the searcher's question directly on the results page, so fewer people ever click through to your site. Organic click-through rate falls by roughly 61% on queries where an AI overview appears, and it keeps falling even on queries where it doesn't, because searchers have learned to expect the answer up top. The fix isn't writing more content. It's writing content that gets cited instead of skipped.

In this article: what an AI overview does to your funnel, the mistake most founders make chasing rank instead of citation, the framework for getting cited, real numbers from a team that lived through this, and the 30-day move to try first.

What an AI overview actually does to your organic traffic

An AI overview is a synthesized answer Google generates at the top of a search result page, built from several sources at once, that lets the searcher get an answer without visiting any of them. This is the mechanism behind what's often called a zero-click search: the query gets answered, no site gets the visit.

Seer Interactive tracked 3,119 informational queries across 42 organizations between June 2024 and September 2025, covering 25 million organic impressions. Organic CTR on queries triggering an AI overview dropped from 1.76% to 0.61%, a 65% decline. Paid CTR fell even further, down 68%. Queries with no AI overview present still saw organic CTR drop 41%, which means the damage isn't contained to the pages Google flags. Searcher behavior changed everywhere.

The part that catches most founders off guard: impressions often go up while clicks go down. Google counts a separate impression for the AI overview slot and the traditional listing below it, so your Search Console graph can show rising visibility at the exact moment your click volume is cratering. One documented case showed impressions up 27.56% year over year while clicks dropped 36% and average position actually improved. Read that dashboard the wrong way and you'll spend a quarter optimizing a metric that no longer means what it used to.

The mistake: still optimizing for rank instead of citation

Most founders respond to a traffic dip by doing more of what used to work: another 2,000-word guide, another round of backlinks, another keyword gap analysis. That playbook assumes the click is still the prize. It isn't.

92.36% of AI overview citations come from domains already ranking in the top 10 organically, so traditional SEO isn't obsolete, it's the entry ticket. But ranking well no longer guarantees the click. Brands that get cited inside the AI overview see 35% more organic clicks and 91% more paid clicks than brands ranking on the same query without a citation. The competition isn't for position one anymore. It's for one of the three to five sources the model decides to name. This is the same shift covered in GEO vs SEO for B2B SaaS: one earns a ranking, the other earns a citation, and you now need both.

Chasing rank while ignoring citation means you can win the old game and still lose the new one: page one, zero clicks, no idea why.

The framework: write to get cited, not just ranked

Getting cited is a different discipline than getting ranked, though it builds on the same foundation. Four things move the needle, in order of leverage.

1. Answer in the first 50 to 70 words. Open every article and every major section with a direct, self-contained answer, written so it makes sense with zero surrounding context. This is the exact chunk a model lifts to build its summary.
2. Cite your own sources. Pages that get cited almost always cite one or two authoritative sources themselves per major section. Models trust content that shows its work.
3. Refresh on a schedule, not on a whim. AI systems weight recent content over older, more comprehensive pages. A mediocre post updated last month can out-cite a great post from two years ago. Prioritize refreshes on pages that used to convert but have gone quiet.
4. Build real topical depth. A single article rarely earns a citation on a competitive query. A cluster of 5 to 8 interlinked pieces covering a topic from every angle (definition, framework, comparison, mistakes, examples) signals the kind of expertise models are trained to trust. Google says as much directly in its own guidance on succeeding in AI search.

Comparison and list-shaped queries deserve their own structure. If the query is "X vs Y," build an actual markdown table. If it's "steps to," build a real numbered list. Models pull tables and lists cleanly; they don't pull them out of a wall of prose.

What this looks like with real numbers

A B2B SEO team running their own blog through this shift found their top-ranked article on this exact topic sat at position 6 across 400,000 monthly impressions, with a CTR of just 0.14%, the lowest on their site despite having the most visibility. Two other articles at a similar ranking position, on topics less likely to trigger an AI overview, converted at three times that rate:

1. AI overviews and CTR topic, position 6.1. 405,000 monthly impressions, 0.14% CTR. This topic always triggers an AI overview.
2. Google core updates topic, position 5.7. 357,000 monthly impressions, 0.48% CTR. This topic sometimes triggers an AI overview.
3. A spreadsheet function guide, position 6.2. 298,000 monthly impressions, 0.29% CTR. This topic often triggers an AI overview.

Same ranking tier, three-times difference in clicks, purely because of how often each topic triggers a synthesized answer.

Citation share is also concentrated. In one analysis of AI mentions across banking-related queries, one large bank held 32.2% visibility across AI platforms for its category, while a smaller, well-positioned competitor still captured outsized mentions relative to its ad spend, purely by being the source models trusted enough to name. Category size didn't decide who got cited. Content structure and authority did.

The industries seeing this hit hardest are exactly the ones full of informational, how-to, and comparison content, which describes most B2B SaaS blogs. If your content answers "what is," "how to," or "X vs Y" questions, you're already in the blast radius whether you've noticed the CTR drop yet or not.

The 30-day move

Pick your 5 highest-impression, lowest-CTR articles from Search Console this week. For each one, rewrite the opening 70 words into a direct, standalone answer, add one authoritative outbound citation, and add a comparison table or numbered list if the query shape calls for it. Then manually test the target query in ChatGPT and Perplexity and note whether you're mentioned, the same way we cover in how to get your B2B SaaS mentioned by ChatGPT. Re-check in 30 days, and if you want to separate this traffic from the rest of your analytics while you test, set up AI referral tracking in GA4 first. This is the fastest, cheapest way to find out whether your content is built to be cited, before you spend another quarter producing more of what isn't working.

Frequently asked questions

Does Google Search Console separate AI overview clicks from regular clicks?

No. As of mid-2025, AI overview and AI mode clicks count toward your existing totals under the "Web" search type, with no separate filter. You can see impressions rise and clicks fall on the same query, but you can't isolate the AI overview's exact contribution from Search Console alone.

Should I block AI crawlers if I'm not seeing traffic from citations?

Generally no. Blocking Googlebot removes you from traditional search entirely, since AI overviews draw from the same crawl. Blocking other AI crawlers removes you from consideration in millions of queries you'd otherwise have a shot at appearing in, even without a click to show for it.

How long before content changes show up in AI citations?

Tactical fixes, direct answers, added citations, fresher data, can show measurable movement in 30 to 45 days. Building durable share of voice across multiple AI platforms takes a full quarter or more of sustained, consistent output.

What is AI share of voice?

AI share of voice is the percentage of times your brand gets named when a model answers questions in your category, tracked across ChatGPT, Perplexity, Gemini, and Google's AI overviews. It is the citation-era equivalent of organic market share.

Is this only a problem for informational content?

It hits informational and comparison queries hardest, which is most B2B SaaS blog content. Commercial, ready-to-buy queries see fewer AI overviews today, though that gap is closing.

The takeaway

The click was never the goal. It was always a proxy for being trusted enough to get chosen. AI overviews just made that proxy stop working, which means the founders who adapt fastest are the ones willing to rebuild their content around citation instead of rank, starting with the five pages bleeding clicks right now. If you'd rather have someone else run that rebuild while you focus on the product, see how costprice.in works.