Search itself is not being reinvented overnight. Search results are.
That is the clearest message from DataForSEO's Search Insider: 2026 Edition. Drawing on more than 190 million keywords, over 165 million Google AI Overview and ChatGPT answers, and roughly 1.1 billion backlink URLs, the report provides a detailed view of how discovery is changing.
People still tend to use relatively concise searches. Yet the page they see after pressing enter now looks very different. AI Overviews are appearing far more often, traditional organic listings are receiving a smaller share of clicks, and visibility increasingly depends on whether a brand appears across search results, AI citations and influential third-party platforms.
Here are the report's most important findings - and what they mean for SEO and generative engine optimisation (GEO).
1. Searches are getting longer, but only gradually
Three- and four-word searches remain the most common, so conventional keyword research is not suddenly obsolete. There is, however, a measurable move towards longer and more specific queries.
Between September 2025 and August 2026, the share of desktop queries containing five or more words rose from 29.3% to 34.8%. On mobile, the increase was smaller, from 29.6% to 31.9%.
This matters because longer searches tend to reveal more context. A user searching for "best CRM" gives little away. Someone searching for "best CRM for a small UK consultancy" signals their use case, market and likely evaluation criteria.
The practical lesson is not to abandon short, high-volume terms. It is to build content that also answers specific questions, comparisons and scenarios. Topic depth, supporting pages and natural-language subheadings can help a site serve both conventional keyword searches and more detailed queries.
2. Top organic rankings are producing fewer desktop clicks
The report's organic click-through-rate data shows a much sharper change than its query-length analysis.
From September 2025 to August 2026, desktop CTR for the number-one organic result fell from 17.83% to 7.43%, a decline of 10.41 percentage points. Position two fell from 15.27% to 3.07%, down 12.20 points.
Mobile proved more resilient. Position-one CTR declined from 25.17% to 20.80%, while position two increased slightly from 7.99% to 8.52%.
DataForSEO is careful not to claim that AI alone caused these changes. Search layouts, user behaviour and other SERP features may all contribute. Even so, the implication is significant: ranking reports can no longer stand in for performance reports. A high position may deliver a very different level of traffic depending on the device, query and surrounding search features.
3. AI Overviews are redistributing the click opportunity
DataForSEO's clickstream-based model estimates that when an AI Overview is present, its citations receive 42-45% of non-zero clicks across informational, navigational, commercial and transactional searches.
At the same time, the estimated share going to the top organic result falls from roughly 33-38% without an AI Overview to 21-24% with one. Positions two to ten also lose a substantial part of their combined click share.
These are modelled estimates rather than directly observed clicks, but the pattern is consistent: an AI Overview does not simply add another feature to the results page. It changes how attention and clicks are distributed.
That makes citation visibility increasingly important. If an AI answer is taking a large share of the available clicks, appearing as one of its sources can become almost as relevant as occupying a traditional organic position beneath it.
4. AI Overviews have moved from occasional to commonplace
Across DataForSEO's historical panel, the proportion of analysed Google results containing an AI Overview increased from 5.93% in September 2024 to 61.43% in August 2026 - more than a tenfold rise.
Over the same period, featured snippets declined from 8.96% to 0.44%. Video results rose from 9.67% to 19.98%, while local packs remained prominent at 25.99%.
The wider point is that Google visibility is becoming more fragmented. Traditional blue links still matter, but they now sit alongside AI-generated answers, video, local results and other formats. SEO strategy therefore needs to consider the full search results page rather than treating the ten organic listings as the complete market.
5. Informational searches lead - but commercial queries are catching up
AI Overviews are most prevalent for informational searches. Their trigger rate rose from 48.42% in Q4 2025 to 70.44% in Q3 2026.
The more commercially important development is the speed of growth elsewhere. Commercial-query trigger rates almost doubled, rising from 26.30% to 51.35%. Navigational searches reached 34.03%, while transactional searches reached 28.64%.
AI competition is therefore no longer confined to the top of the funnel. It is extending into comparison, brand and purchase-oriented journeys. Forecasts based only on search volume and expected organic CTR may overstate the traffic available unless they also account for AI Overview exposure.
6. Search volume, difficulty and CPC do not provide a safe haven
Several of the report's findings challenge familiar keyword-selection assumptions.
Keywords with 201-300 monthly searches triggered AI Overviews 58.59% of the time in Q3 2026, compared with 48.62% for keywords with more than 600 searches.
Mid-difficulty keywords, with difficulty scores between 21 and 60, produced AI Overviews around 65% of the time. The most difficult group, scoring 81-100, had a lower rate of 53.76%.
Keywords costing $5 or more per click triggered AI Overviews 49.55% of the time, close to the 50.54% rate for keywords costing up to $0.25.
In other words, low-volume, relatively achievable or expensive keywords are not automatically protected from AI-driven competition. Keyword research now needs another layer: the current and likely shape of the results page.
7. AI Overview exposure varies dramatically by industry
The average figures conceal large differences between sectors.
Health had the highest AI Overview trigger rate at 69.68%, followed by Finance at 68.80% and Family & Community at 68.63%. At the other end of the spectrum, Real Estate stood at 22.27%, Dining & Nightlife at 29.37%, and Apparel at 37.48%.
That spread of more than 47 percentage points means there is no universal answer to how urgently a business should adapt. A healthcare or financial-services brand may already operate in an AI-heavy search environment, while a real-estate business may still see a more conventional results page.
Benchmarking and forecasting should therefore be sector-specific - and ideally based on the actual queries that matter to the business.
8. Google AI Overviews and ChatGPT favour different sources
Visibility in one AI platform does not guarantee visibility in another.
YouTube was the most-cited domain in Google's AI Overviews, with 51.9 million citations in the analysed snapshot. Reddit followed with 25.1 million, then Facebook with 20.4 million and Wikipedia with 19.9 million.
ChatGPT showed a different pattern. Wikipedia led with 4.88 million citing answers, followed by Reddit with 3.25 million. Reference, health and publishing sites featured more prominently, while several social platforms in Google's top ten did not appear in ChatGPT's top 30.
The two systems also differed in how often they provided sources. The report found that 27.3% of analysed ChatGPT answers had no citation, compared with 0.7% of Google AI Overview answers.
For marketers, this makes GEO a platform-specific discipline. The sources, formats and authority signals associated with visibility on Google cannot simply be assumed to work identically in ChatGPT.
9. Reddit is now part of the mainstream search landscape
Reddit appeared in Google's top ten for 32.15% of the analysed search results in Q3 2026. Its presence remained close to one-third across four consecutive quarters and was almost eight times Quora's 4.11% rate.
Reddit was also the second-most-cited domain in both Google AI Overviews and ChatGPT within DataForSEO's datasets.
This does not mean brands should flood communities with promotional posts. It means genuine customer conversations can influence discovery well beyond the platform itself. Search teams should know which discussions rank for their priority topics, what questions people ask, which brands users recommend and whether useful first-hand expertise is missing.
Community visibility is becoming an important part of search visibility.
10. Backlinks still matter, but correlation is not causation
The report finds a moderate positive relationship between backlinks and AI citations. UGC backlinks correlated with citations in both Google AI Overviews and ChatGPT, but the number of referring domains overall had a stronger association.
The traditional organic pattern also remains visible. Domains ranking first had a median of 111 referring domains, compared with 85 for domains ranking in positions four to ten - a 31% difference.
For Google AI Overview citations, the median backlink profile rose with keyword difficulty. Domains cited for keywords with difficulty scores of 81-100 had 1,571 referring domains at the median, 5.6 times the 282 median for keywords scoring 1-20.
DataForSEO repeatedly stresses that these are associations, not proof that links directly caused rankings or citations. The sensible takeaway is not to chase UGC links or raw volume. It is to keep building a diverse, relevant and credible web presence that earns references from appropriate sources.
What should search marketers do next?
The report points towards a broader model of discoverability. Five actions stand out:
Audit visibility beyond your website. Identify the publishers, communities, videos and reference sites that rank or receive citations for your important topics.
Analyse AI exposure at keyword level. Add AI Overview presence and citation sources to keyword research, traffic forecasts and competitor reviews.
Treat Google and ChatGPT separately. Measure each platform's results, sources and citation patterns instead of grouping everything under one GEO score.
Strengthen authority across the web. Prioritise useful content, relevant digital PR, expert contributions and a varied referring-domain profile over link-count shortcuts.
Report on multiple surfaces. Bring organic rankings, AI citations, third-party mentions and community visibility into the same measurement framework.
The bottom line
DataForSEO's research does not suggest that SEO is being replaced. It suggests that SEO's boundaries are expanding.
Websites and backlinks remain important, but a brand can now be discovered through an AI citation, a Reddit recommendation, a YouTube video or a publisher's article before a user ever reaches its own site. The job is no longer only to rank a page. It is to build a credible presence across the sources and formats that search engines and AI systems use to construct answers.
For search teams, that means keeping the fundamentals - technically sound websites, useful content and genuine authority - while measuring a much wider field of competition. In 2026, visibility is not a single position. It is an ecosystem.
Ready for Search Everywhere Optimisation?
At Black Crow Digital, we call the next phase of organic growth Search Everywhere Optimisation (SEvO): making your brand visible, credible and useful wherever people discover, research and compare - from Google and AI answers to YouTube, social platforms, communities and trusted publishers.
SEvO is where our focus is moving. It brings together the foundations of SEO with AI visibility, digital PR, content, community and brand authority, helping businesses become the trusted answer across the whole search ecosystem.
Want to see where your brand is visible - and where the gaps are? Talk to Black Crow Digital about building your SEvO strategy.
Source and methodology note: This article summarises DataForSEO's October 2026 report, “Search Insider: 2026 Edition”. Its primary search and AI analyses focus on US, English-language queries and responses; backlink data has global coverage. Dataset periods and samples vary by research question. Click-share figures are modelled estimates, and reported backlink relationships are correlations rather than evidence of causation.