Popular Posts

Latest Articles

Semrush Partner

What AI Overviews Did to Search in Türkiye

Google’s AI Overviews and AI Mode went live in Türkiye on 18 February 2026. Four and a half months in, we stopped guessing and measured.

Google's AI Overviews and AI Mode went live in Türkiye on 18 February 2026. Four and a half months in, we stopped guessing and measured.

Google’s AI Overviews and AI Mode went live in Türkiye on 18 February 2026. Four and a half months in, we stopped guessing and measured.

I didn’t lean on a single source. I combined three: the visibility shift of each sector’s market leaders (Semrush, publicly available), 1,035 live SERP queries across 17 sectors (about 9,740 organic results), and the first-party Search Console data of the sites we track.

The result is clear: rankings held, but clicks on informational queries lost a third. The hardest hit landed on finance and insurance (CTR down ~60%). Transactional verticals like travel fell more slowly, but they still fell (market leaders −8%).

One more striking finding: high Domain Authority does not guarantee you a spot in the AI Overview. I poured all of it into a live, bilingual dashboard that I’ll refresh every month.

Hey everyone,

For months we all repeated the same line: “AI Overviews are coming, they’ll change search.” Back in February I even said, on video, that this was “the biggest shift in twenty years.” Well, on 18 February 2026 Google switched on AI Overviews and AI Mode in Türkiye. Four and a half months have passed. So this week I’m not reporting news. I stopped guessing and measured it.

I combined three data sources: the visibility shift of each sector’s market leaders (Semrush, publicly available), 1,035 live SERP queries across 17 sectors (about 9,740 organic results), and the first-party Search Console data of the sites we track. Then I poured all of it into a live, bilingual dashboard that I’ll refresh every month: turkiye-yapay-zeka-aramasi-izleme.netlify.app. Flip the language toggle top-right for English.

Grab a coffee. Let’s read the data together.

The core finding: rankings held, clicks eroded

Here’s the whole story in one sentence: positions barely moved, but on informational queries clicks fell by a third. Across the sites we track, total clicks per day were down 33% in June versus the pre-launch baseline, while average position held around 7.6. That is the classic AI Overview fingerprint. The user gets the answer in the summary box and never clicks the blue link. You’re not losing rankings; the click is evaporating. And these two things are completely different, which is exactly what most people confuse.

The market leaders tell the same story. In health, the top three players together lost 38% of their combined organic visibility in the launch month alone, and 56% by June. Travel fell more slowly, but it still fell.

What I actually did (method, in the open)

A number is only as good as the way it was collected, so here it is:

  • Sample: sector market leaders (Semrush, public), 10 sites we track (aggregated, no client names), and 1,035 representative live SERP queries (17 sectors, ~9,740 organic results).
  • Window: 1 Feb – 30 Jun 2026. The comparison baseline is the period before AI Overviews launched on 18 February (1–17 Feb). Monthly trend covers March–June.
  • Sources: Google Search Console (clicks, impressions, CTR, position), Ubersuggest live SERP (AIO trigger, SERP features, Domain Authority), and Semrush (market leaders and monthly organic traffic trend).
  • The metric: my cleanest AIO signal is “CTR falling while position holds.” If ranking is flat but CTR drops, the summary box is most likely eating the difference.
  • The honest caveat: Feb to June runs winter into summer, so seasonality is a factor, not just AI Overviews. Semrush market-leader traffic is a modeled estimate. For finance/insurance, where the Semrush estimate is too volatile, I used first-party GSC CTR instead. Travel is read from market-leader data, not a single seasonal site.

Sector by sector

The two ends of the spectrum split apart.

Finance and insurance took the hardest hit. CTR here dropped about 60% versus the pre-launch baseline. Impressions rose, so the content showed up more, but the clicks collapsed. Questions like “how do I raise my credit score” are exactly what an AI Overview answers in one paragraph.

Education was hit hard too, with CTR down around 36%. Exam, application and study-abroad queries are perfect AIO fodder.

Tech media bled clicks steadily, roughly 23–26% month over month. Position holds, clicks leave.

E-commerce split in two. The content-heavy marketplace side lost clicks (around 34%), while the pure transactional B2B side barely moved. “Best refrigerator” triggers an AIO; “refrigerator prices” does not.

Travel fell slowest, but it still fell. The market leaders together slipped about 8%. Transactional intent protects it a little, but there’s no growth.

Which queries trigger an AI Overview?

Across 17 sectors I checked 1,035 queries live. An AI Overview appeared on 59% of them, but the spread is wildly uneven: highest in parenting (90%) and beauty (83%); tech, education and health at 66–67%; finance 62%; e-commerce 56%; lowest in travel (26%). By intent the split is sharpest of all: 70% on informational queries, 37% on transactional, and just 25% on tool/calculator queries. So AIO isn’t everywhere; it concentrates on informational and comparison intent. That matches the global picture almost exactly.

What does AI cite? Not authority. Relevance and community.

Does high Domain Authority get you into the AI answer? Good question, and the honest answer comes from large-scale work. Peec AI, analyzing 30 million sources across five engines, found Reddit is the single most-cited source; Google’s AI Mode and AI Overviews favor social and community content. OtterlyAI found 41% of cited YouTube videos had under 1,000 views. So the deciding factor isn’t raw authority or popularity; it’s relevance, structure and brand mentions.

I measured the same thing at scale this week: 1,035 live queries, 9,740 organic results. The result points the same way as the big studies. On SERPs where an AIO appears, average DA is not higher; it’s slightly lower (44.6 vs 47), and the share of low-DA pages is higher (32% vs 27%). What actually rises is community. Our local “Reddit,” Ekşi Sözlük, appears in 37% of AIO SERPs versus only 19% of non-AIO ones; forums 21% vs 11%; Reddit 15% vs 6%; YouTube 49% vs 40%. In short: high DA doesn’t get you into the AI answer; relevance, structure and community do.

Who gets cited: your local community platform

Globally, the number-one source for AI search is Reddit. In Türkiye, our Reddit is Ekşi Sözlük. The lesson generalizes: your community strategy should be built on whatever community platform your market actually uses: Reddit in the US, Ekşi in Türkiye. YouTube rises in both markets: on commercial reviews in the US, on health queries in Türkiye.

18 February was a start, not a finish

When AI search launches in a country, the impact doesn’t arrive all at once; it spreads over months. Public studies put AIO coverage at roughly 6% in January 2025, rising to an estimated 48–65% in the US by early 2026 (Google says “~50%,” trackers like Xponent21/AWR find 60–65%). Our own data deepened month over month too: total click loss went from 17% in March to 33% in June; finance CTR decline widened from 48% in March to 60% in June. And in some sectors CTR recovered a little in June, meaning content is starting to adapt, which matches Seer Interactive’s “trough, then partial recovery” pattern across 53 brands and 5.47M queries. So for Türkiye this isn’t over. It’s just beginning, and we’re near the start of the curve.

So where does the traffic go? It doesn’t vanish, it redistributes

Do these clicks disappear entirely? No. Two trends stand out, both backed by data: a shift from generic to branded search, and growth in referral traffic from AI engines to sites.

Per Digital Applied’s March 2026 data, branded-query CTR rises 18% when an AIO shows, and brands cited inside an AIO get 35% more organic clicks. The mechanism: when ChatGPT or an AI Overview recommends a brand, some users search that brand on Google instead of clicking the link. So generic/informational clicks fall while branded searches rise.

The other side is strengthening too: referral traffic from AI engines is growing fast. Per Similarweb, ChatGPT’s web visits grew 84% from September 2024 to March 2026; that traffic is still around 1% of the web, but it converts far better than organic (ChatGPT referral around 7%, organic around 2%). So the picture isn’t a collapse; it’s a reallocation: volume down, intent and brand value up. The strategy shouldn’t be “rescue the click count,” it should be “become the brand the AI cites and people search for.”

The user dropped the keyword. Did your content?

Underneath all of this is a deeper shift: users no longer type keywords, they talk. Per Google’s “How People Are Using AI Mode,” the average AI Mode query is three times longer than classic search. People describe context in the search bar: “I have flat feet and my knees hurt, recommend a running shoe that won’t make it worse.” Follow-up questions are growing more than 40% month over month. One in six searches is now multimodal (voice, image, video), and “which”-style queries grew 40% faster over six months. Take your top ten pages, rewrite each page’s core keyword into the natural sentence an AI Mode user would actually ask. If your content doesn’t answer that longer question, a competitor who does will fill the gap.

For executives: “ranking” no longer guarantees traffic

One sentence for founders and marketing leaders: if you’re still asking “what position are we in,” you may be missing that the rules changed. The same data proves it. Position held (around 7.6) but clicks fell by a third; the ranking held, the traffic left. On 59% of queries the answer sits above the blue link, at the top of the page. The brand cited inside the AIO gets 35% more clicks, while even an uncited #1 loses ground. And what gets into the summary isn’t high Domain Authority.

So the new question isn’t “what’s my rank,” it’s “is the AI citing me, am I inside the answer, is my brand being searched.” Ranking is now an indicator, not the outcome. Team goals and reporting need to shift accordingly: put visibility, citations and branded search next to rank tracking.

What to do about it: GEO and AEO moves

Concrete steps from this data:

  • Put the answer first (BLUF). Give a clear answer in the first 150 words. AIO takes its answer from the top of the page.
  • Firm up schema and E-E-A-T. FAQ, Article, Author schema; author identity and dates. This is the plumbing of citability.
  • Invest in YouTube. Timestamped, long-form video. AIO and AI Mode citation is strong here, especially in health.
  • Show up in your local community platform. In the US that’s Reddit; know your market’s equivalent.
  • Build authority by topic, not by DA. High DA doesn’t guarantee entry into AIO; topic depth and structure do.
  • Protect transactional and local pages. AIO doesn’t trigger on these; real traffic and conversion still live there.
  • Measure on the right axis. Watch CTR decline at a stable position, not the ranking itself. That’s where the real AIO effect shows.
  • Be agent-ready technically. Allow AI bots in robots.txt, keep clean text structure, evaluate an llms.txt.

The dashboard: public, monthly, bilingual

I didn’t leave this as a one-off post. It’s a live, bilingual dashboard with monthly click and CTR trends, the sector breakdown, AIO trigger rates, the Domain Authority analysis, the Türkiye-vs-world citation comparison, and the full method. Toggle between English and Turkish top-right.

👉 Explore the dashboard here: turkiye-yapay-zeka-aramasi-izleme.netlify.app

And it updates every month, because as we’ve seen, this effect isn’t static, it deepens. Next month’s data will show where the curve is heading, and we’ll watch it together.

Share!

These May Also Interest You

Craving more SEO knowledge? Extend your learning with #SEOSDINERSCLUB

Subscribe to our newsletter for weekly SEO insights, join the discussion in our community, or engage with professionals on our Twitter group.