ChatGPT
Claude
Perplexity
Conductor’s AEO and GEO Benchmarks Report offers data behind what’s happening to search as AI becomes part of the experience. It looks at how people discover content through LLMs, how Google is answering more queries in AI Overviews, and what sources are cited by AI.
The dataset covers 13,770 domains across multiple industries and search environments. Conductor compares AI referral traffic from tools like ChatGPT, Gemini, and Copilot with visibility patterns in Google AI Overviews, then breaks the results out by vertical to narrow down the impact.
I’m pulling the key data from the report and breaking down what it actually means: how big AI referrals really are, how often AI Overviews show up, where citations tend to come from, and if AI search is worth optimizing for.
Summary of Key Findings
- LLMs account for around 1.08% of total site traffic on average, with variation by industry.
- AI referral traffic is growing at approximately 1% month over month, but growth patterns are uneven and not consistent across verticals.
- ChatGPT drives roughly 87% of AI referrals overall, though other LLMs such as Gemini and Copilot have larger shares in specific industries.
- Google AI Overviews appear on about 25% of queries, with higher exposure in some industries (48.7% in Healthcare) and lower exposure in others (4.4% in Real Estate).
- AI Overviews cite familiar content formats and pull from different source types by industry, including official documentation, publishers, platforms, and institutions.
- Brand mentions and source citations don’t always overlap; brands may be referenced in AI answers without being the cited source.
- Clean attribution of Google AI experiences isn’t possible using standard GA reporting, as AI channels blend into organic traffic.
AI Referral Traffic is a Small Percentage – But Growing
Conductor’s data shows that traffic coming directly from AI is currently just over 1% of total site traffic on average across different industries. This includes referrals from LLMs like ChatGPT, Gemini, and Copilot. It’s measured separately from traditional organic search.
That overall average covers a wide spread by industry. IT has the highest share of AI referral traffic at roughly 2.8%, followed by Consumer Staples at around 1.9%. Both sit well above the mean, which lines up with sectors where people spend more time researching products, tools, or technical information before taking action.
On the low end, AI referrals are close to negligible. Communication Services average around 0.25%, and Utilities are closer to 0.35%. In those verticals, AI accounts for a very small fraction of visits even when you look across the full reporting period.
The real story here is the percentage overall. Even in the best-performing industries, AI referrals are still a small slice of total visits. The data supports two things: AI is sending measurable traffic, but it hasn’t become a primary traffic source to rely upon. The difference between 0.25% and 2.8% is greater when you consider it in the millions, but both numbers are still far below what you’d associate with a core acquisition channel.
AI is producing referral traffic, it’s unevenly distributed, and it’s trackable. It’s just limited right now when you compare it to overall site traffic. A sign that GEO/AEO strategies shouldn’t be replacing SEO, but become a small addition to the overall strategy.
Organic Google Search is Still the Primary Search Channel
The report compares AI referral traffic directly against traditional traffic sources to show how it fits into the overall acquisition mix. Across every industry analyzed, AI referrals represent the smallest share of total website traffic, while established channels continue to account for the majority of visits.
Organic Google traffic alone consistently represents a large portion of total traffic. In Healthcare, organic search accounts for 42.36% of average website traffic. Communication Services follows at 39.65%, and Industrials at 33.85%. Even in industries with lower organic reliance, organic search still contributes a larger share of traffic than AI referrals.
By comparison, AI referral traffic is below 3% across all industries, peaking at 2.80% in Information Technology and 1.91% in Consumer Staples, and falling as low as 0.25% in Communication Services and 0.35% in Utilities. These place AI referrals well below not only organic search, but also direct, paid, social, and other traditional referral channels.
Other non-AI channels dominate total traffic share in most verticals. In Consumer Staples, Real Estate, and Utilities, direct, paid, social, and traditional referral traffic collectively account for over 80% of total visits, with some industries approaching 90%. Against that baseline, AI referrals are marginal in absolute volume.
This comparison shows that AI referral traffic works as a supplementary part of existing channels rather than displacing them. Organic search is still a primary traffic driver across industries, and the overall traffic mix continues to be defined by established acquisition sources, with AI referrals representing a small addition rather than a replacement.
AI Referrals Are Growing, But It’s Inconsistent
The report shows that AI referral traffic is increasing over time, but not at a uniform rate. When aggregated across all industries, AI referrals grow at roughly 1% month over month on average, indicating gradual expansion rather than sudden acceleration.
That overall trend breaks down quickly at the industry level. Industries that already have higher baseline AI referral traffic, such as IT and Consumer Staples, tend to show more consistent month-over-month increases. In these categories, AI referrals rise incrementally across consecutive months rather than appearing as isolated spikes, suggesting a steadier pattern of use.
In contrast, industries with very low baseline AI referral traffic show flat or unstable growth patterns. Utilities and Communication Services, which are already at the bottom of AI referral share, also display irregular movement over time, with some months showing slight increases and others showing no change. In these cases, AI referrals are marginal even with short-term growth.
The report also shows that growth rates are influenced by starting position. Industries beginning at fractions of a percent can register month-over-month growth without materially changing their overall contribution to traffic. This makes percentage growth figures difficult to interpret without considering absolute scale alongside them.
Taken together, the data shows that AI referral growth fluctuates. Aggregate averages suggest a gradual increase, while industry-level data show a mix of steady growth, flat performance, and volatility. The report documents movement, but not a single, consistent growth pattern across markets.
ChatGPT Drives Most AI Referrals, But Not in Every Industry
Across all industries analyzed, ChatGPT accounts for roughly 87% of AI-referred traffic on average. At an aggregate level, this is why AI referral traffic is often discussed as if it comes mainly from ChatGPT.
However, the report also shows that this distribution isn’t consistent across industries. While ChatGPT is the largest single source in every vertical measured, the share attributed to other LLMs increases in certain categories.
In Utilities, for example, Gemini accounts for approximately 21% of AI referral traffic, a much higher share than the cross-industry average. This indicates that Google AI experiences play a larger part in discovery for this sector than they do elsewhere.
In Financial Services, Copilot contributes up to around 5% of AI referrals, a smaller absolute share, but still notable given the overall size of AI traffic in that category.
These differences show that AI referral traffic isn’t a single, uniform channel. Even when total AI referrals are small in absolute terms, the mix of LLMs sending that traffic varies by industry. A traffic source that is marginal in one vertical may give a larger portion of AI referrals in another.
The data shows two things: ChatGPT is the dominant source of AI referrals overall, but it isn’t the only contributor, and its share isn’t fixed. AI referral traffic is fragmented across LLMs in ways that depend on industry context rather than following a single global pattern.
Google AI Overviews Appear More, but Send Less Traffic
While LLMs send a small share of referral traffic, Google AI Overviews appear on a much larger portion of searches. Across all industries analyzed, the report shows that approximately 25% of queries trigger an AI Overview. This makes AI Overviews a more common point of exposure than AI referral traffic when measured at the query level.
That overall figure varies by industry. Healthcare shows the highest AI Overview presence, appearing on around 48.7% of queries. This means nearly one out of every two health-related searches in the dataset triggers an AI answer directly. Financial Services and Utilities sit closer to the overall average, with AI Overviews appearing on roughly 25.7% and 25.4% of queries respectively.
At the lower end of the range, AI Overviews are less common. Real Estate queries trigger AI Overviews at approximately 4.4%, while Consumer Staples sit around 6.8%. In these categories, AI answers are the exception rather than the norm, at least within the scope of queries analyzed.
AI Overviews appear on roughly one quarter of searches overall, while LLMs account for around 1% of site traffic on average. These are different points measuring different behaviors.
AI Overviews show on-SERP visibility, where information is given directly to the user, while AI referrals represent off-SERP clicks, where a user leaves the LLM or search experience to visit a site.
The report doesn’t measure click impact directly from AI Overviews, but it does make clear that frequency of appearance and volume of traffic are not the same thing. AI Overviews are widespread across certain industries, even where AI referrals remain limited, which positions them as important for visibility rather than a traffic channel.
AI Overviews Cite From Common Page Types
The report identifies a clear and limited set of page types that AI Overviews pull from across industries. Rather than citing new or specialized formats, AI Overviews consistently reference existing, widely published content types.
The report lists the most commonly cited page types as:
- Blog posts
- Editorial articles
- News content
- Videos
- Product or service pages
These formats appear across the industry breakdowns included in the analysis. While the report doesn’t assign percentages or frequency weights to each format, it does show that AI Overviews rely on a specific and familiar range of page structures, rather than niche or experimental content types.
The report also shows no meaningful variation in cited page types by industry. Although the sources of those pages differ, for example, documentation-heavy sites in IT or institutional publishers in Healthcare, the underlying formats are consistent. Articles are cited as articles, videos are cited as videos, and product pages are cited where product-level information is relevant.
What the data makes clear is that AI Overviews are not built from a distinct category of ‘AI-first’ content. Instead, they reuse standard web formats that already dominate publishing across industries. This shows that it’s less about creating new content types and more about how existing pages get cited.
Importantly, the absence of numeric weighting here is part of the finding. The report doesn’t suggest that one format materially outweighs others across all contexts, only that AI Overviews repeatedly draw from the same small set of established page types.
Brand Citations Vary by Industry
Across industries, AI visibility concentrates around a small set of leaders rather than being spread evenly. This was measured by ranking market share leaders based on domain citation share and brand mention share across 17 million AI-generated responses and 100 million citations.
The report makes a clear distinction between brand mentions and citations, and shows that the two do not always overlap. A brand can appear frequently in AI answers without being the source that AI actually cites.
This is a breakdown of domain citations and brand mentions by industry:
Communication Services
Citations: YouTube, Reddit, Google, Investopedia, Reuters
Brand mentions: Google, YouTube, Investopedia, Roku, Forbes
Consumer Discretionary
Citations: Clemson University, Walmart, Target, Four Seasons, Cornell University
Brand mentions: Walmart, Target, Miele, Sonos, Nike
Consumer Staples
Citations: Amazon, Walmart, Chewy, Best Buy, Mattress Firm
Brand mentions: Amazon, Walmart, Best Buy, Mattress Firm, Chewy
Financial Services
Citations: NerdWallet, Bankrate, Kiplinger, Vanguard, Experian
Brand mentions: NerdWallet, PayPal, Bankrate, Vanguard, Fidelity
Healthcare
Citations: Mayo Clinic, Cleveland Clinic, Healthline, GoodRx, WebMD
Brand mentions: Mayo Clinic, Cleveland Clinic, Pfizer, Labcorp, Baptist Health
Industrials
Citations: Deloitte, Indeed, McKinsey, Wolters Kluwer, Siemens
Brand mentions: Deloitte, Amazon Web Services, Siemens, ADP, McKinsey
Information Technology
Citations: Google, Microsoft, SAP, Dell, Adobe
Brand mentions: Google, Microsoft, SAP, Apple, Adobe
Materials
Citations: Evonik, Airgas, Trex, SpecialChem, Ingredion
Brand mentions: Trex, 3M, Evonik, Airgas, Ingredion
Real Estate
Citations: Hines, Public Storage, CBRE, Extra Space Storage, Colliers
Brand mentions: Hines, Public Storage, CBRE, Zillow, Colliers
Utilities
Citations: New Fortress Energy, Denver Water, Constellation, EnergySage, GE Vernova
Brand mentions: New Fortress Energy, GE Vernova, Constellation, Denver Water
Taken together, these rankings show that AI visibility is highly concentrated and strongly influenced by industry context. Leadership is split between platforms, publishers, institutions, and large enterprises, depending on the vertical, and the brands with the most AI exposure are not the same across markets.
There’s also a clear split between who gets cited and who gets talked about. Some industries are dominated by platform domains in citations (like YouTube and Reddit), which can mask the underlying brand that actually created the content.
In others, citations and brand mentions line up closely (like Consumer Staples, where Amazon and Walmart show up in both lists), which is a sign that domain authority and brand recognition are reinforcing each other.
The result is that ‘AI visibility’ isn’t one thing; it’s a mix of citation share and mention share, and the balance between the two changes by vertical.
What the GEO/AEO Data Shows Together
Across the report, AI referrals, AI Overviews, and AI citations are all part of discovery, but they work differently and at different scales. Referral traffic from LLMs is measurable but is still small compared to overall site traffic. Google AI Overviews appear on a higher share of queries in some industries, which increases visibility without necessarily producing clicks.
The data also shows that AI visibility is uneven. Both referrals and citations vary by industry, and market share is concentrated rather than evenly distributed. A small number of domains and brands account for a large share of AI exposure within each vertical, while most sites see little to none. The leading brands in one industry are not the leaders in another, and high visibility in one context doesn’t automatically relate across markets.
Taken together, the data shows you can’t treat AI as one channel. Referrals, Overviews, and citations move differently, and the winners change by industry. This report is useful because it shows those differences instead of flattening them into one metric.
What the Data Doesn’t Prove
The report provides benchmarks across industries, but it doesn’t support several claims that tend to get layered onto ‘AI search‘ discussions.
First, it doesn’t show AI replacing search as a primary acquisition channel. AI referrals average about 1.08% of total traffic, while AI Overviews appear on about 25.11% of queries. That describes low click volume alongside higher on-SERP visibility, not a change in demand away from SEO into LLMs.
Second, it doesn’t prove AEO or GEO ROI by industry. The report shows where AI visibility is concentrated and how it varies by vertical, but it doesn’t connect that directly to revenue outcomes in a way that can be generalized across markets. It isn’t an ROI model, and it doesn’t provide industry-level performance benchmarks for business impact.
Third, it doesn’t establish long-term trends. The analysis covers a specific window and shows modest movement (including 1% month-over-month AI referral growth), but it doesn’t establish whether the same patterns hold over longer timeframes or across changing search features and model behavior.
