TL;DR
- A Semrush study published 8 September 2026 found AI assistants and AI Mode sent 0.48% of US sessions to manufacturing and industrial sites from January to July 2026.
- AI Overviews grew from 38% to 57% of tracked manufacturing search volume in those seven months, reaching part numbers and branded queries.
- Direct traffic was 56.65% of sessions and organic search 22.28%, with direct share rising as organic share fell.
- Only Vevor and Grainger were in the top 15 for both AI mentions and AI citations; Semrush sells AI visibility tools and calls the analysis directional.
AI assistants and Google's AI Mode together sent just 0.48% of all sessions to manufacturing and industrial websites between January and July 2026, according to a Semrush study published on 8 September 2026 and reported by PPC Land. Over the same seven months, AI Overviews spread from 38% to 57% of the sector's tracked search volume. The study's sharper finding is that the brands AI answers mention most are rarely the websites AI answers link to: of the top 15 domains on each list, only two appear on both.
Semrush sells AI visibility tracking and closes the study by pointing readers to those products, so these are a vendor's measurements of a market it serves. Semrush itself calls the combined analysis directional rather than definitive. With that caveat stated, the data separates two things marketers often treat as one, being named and being linked, and that distinction applies well beyond manufacturing.
How the study was built
The research, credited to Luke Harsel with contributor Nikola Manojlović, combines three Semrush datasets over January to July 2026.
- Keyword and results-page data. Semrush took the top 20 manufacturing domains from its Trending Websites ranking, pulled each one's top 1,000 keywords, and kept only keywords where at least two of those domains held a top ranking. That left 458 keywords, tracked monthly and weighted by search volume.
- Clickstream traffic. United States sessions to sites in ten categories: chemicals, civil engineering, construction and maintenance, equipment and supplies, machinery, manufacturing, plastics and polymers, renewable energy, transportation and logistics, and warehousing.
- AI visibility data. The US manufacturing and industrial vertical across ChatGPT, Gemini, Google AI Mode and Google AI Overviews, capturing the top 25 domains by mentions and the top 25 by citations each month.
AI Overviews reached further down the funnel
Across the 458 keywords, AI Overviews grew from 38% of search volume in January to 57% in July. Semrush notes the feature is no longer limited to broad questions. Terms that gained an AI Overview during the window include the product query "pneumatic cylinder", the part number "csla2gf" and the branded search "National Instruments". A part number sits about as far down the funnel as an industrial search goes, and a branded query is the one a supplier expects to own.
Where the traffic actually came from
The clickstream data gave this share of sessions for the ten industrial categories between January and July 2026:
| Channel | Share of sessions |
|---|---|
| Direct | 56.65% |
| Organic search | 22.28% |
| Referral | 12.89% |
| 3.52% | |
| Paid search | 2.58% |
| AI assistants | 0.45% |
| AI Mode | 0.03% |
Organic social (1.33%), paid social (0.19%) and display (0.06%) make up the rest. Direct and organic search together account for 78.93% of sessions. The two AI channels combine to 0.48%.
Within that stable-looking mix, one trend is consistent: direct traffic share is growing while organic search share declines. Semrush says demand for industrial goods has not visibly shifted. The route people take to reach a supplier has.
The "dark SEO funnel"
This is the finding with budget consequences. Organic search is measurable by query and attributable to a page. Direct traffic is the bucket that absorbs sessions analytics cannot classify, including a visit that began with a chatbot conversation and ended with someone typing a brand name into the browser. Semrush cites its own survey of more than 600 business-to-business professionals, which found buyers using AI tools to build vendor shortlists for purchases over $1,000 and then following up with a Google search or a direct visit. A marketer quoted in the study calls this the "dark SEO funnel": rising direct traffic or branded search may be AI visibility working, with no AI referral ever appearing in analytics.
Mentions and citations are two different leaderboards
The study separates two metrics. A mention is a brand name in the text of an AI answer. A citation is a link the system offers as a source. Of the top 15 domains on each list, only Vevor and Grainger appear on both.
Figures Semrush circulated to press on 9 September put Boeing at 771,000 mentions across the four AI systems, ahead of 3M at 743,000 and John Deere at 732,000, with Boeing also leading on Google AI Overviews alone at 286,000. Those counts are not in the published study text, which names 3M, John Deere and Boeing as the top-mentioned brands in a different order; PPC Land reported both rather than resolving the mismatch. What is not disputed is that Boeing does not appear among the most-cited sources at all.
The 15 most-cited domains are mostly directories, marketplaces, distributors and reference publishers. Manufacturers hold four of the fifteen places. Semrush's explanation is structural: mentions reward recognition, and many top-mentioned brands are around a century old, while citations reward the page that best answers the question, which is rarely a manufacturer's own site.
Authority tracks mentions more than citations
Semrush compared its Authority Score across both groups. Top-mentioned brands scored consistently. Top-cited sources did not: engineerfix.com, the single most-cited domain in the sector, has a score of 17, while made-in-china.com sits at 77. The study reads that spread as evidence that site authority lines up with mentions more closely than with citations.
One site ChatGPT cites heavily and Google barely ranks
The clearest platform split involves engineerfix.com, which leads ChatGPT citations but does not reach the top-cited list on Gemini, AI Mode or AI Overviews. Semrush data puts it at roughly 3,700 ranking keywords on Google with little estimated Google traffic, yet close to 300,000 estimated monthly organic visits. The difference is the referral mix: DuckDuckGo, Bing and Yahoo account for more than 68% of its traffic, Google for under 2%. A site that is weak on one search engine can be a strong source on others, and an assistant grounding its answers in those indexes inherits that view. A strategy that wins citations on one assistant carries no guarantee on another.
Citations do not convert into visits at a predictable rate
Matching AI traffic against mentions and citations, the study found no clean correlation. Engineerfix leads on citations yet collects a fraction of the AI referrals lower-ranked domains receive. Made-in-china.com and rsc.org are among the highest AI-traffic earners on the cited list.
Semrush's explanation is what a page offers after the answer ends. Engineerfix publishes broad guides and FAQ-style explanations an AI system can summarise completely, leaving no reason to click. Grainger is a distributor where a citation lands on a page to order from. The Royal Society of Chemistry shows the split inside one domain: its most-cited pages are periodic-table entries in an /element/ folder, while its most-clicked pages are peer-reviewed papers in a /content/ folder. Citations land on reference material; clicks land on the research.
What the source did not say
The traffic data covers United States sessions only. The study does not give conversion or revenue data by channel, so it cannot say whether AI-referred visits are worth more or less than organic ones. It does not explain the mismatch between the published mention order and the press figures. It measures ten industrial categories, not all B2B. And as a vendor study, its methodology has not been independently replicated.
What to check in your own analytics
- Look at the direct share trend, not only AI referrals. If direct and branded search are rising while organic declines, AI discovery may be part of the reason even with near-zero AI referrals.
- Track mentions and citations separately. Being named in an answer builds recognition; being linked sends visits. They need different work.
- Audit which of your pages AI can fully absorb. Pages that answer completely in a paragraph earn citations without clicks. Pages with product data, pricing requests, tools or original research give a reason to visit.
- Check non-Google search engines. Bing-heavy visibility can influence assistants that ground answers outside Google.
- Configure AI referral channels in analytics. A custom channel group in GA4 for known AI assistant referrers makes the small number visible and trackable.
What this means for Thai marketers
The data is US-only and says nothing about Thai buyers. For Thai manufacturers, distributors and B2B service firms, the structure of the finding still applies. Industrial buyers in Thailand also shortlist suppliers through search and, increasingly, AI assistants, and directories and marketplaces can hold the citation even when a manufacturer holds the brand recognition. Setting up GA4 tracking that separates AI referrals and watches the direct channel, building product and specification pages that give a reason to click, and developing a GEO plan for being cited rather than only mentioned, are sensible steps. An SEO Thailand program that covers both Google and Bing visibility addresses the platform split the study describes.
Frequently asked questions
How much manufacturing website traffic comes from AI?
According to Semrush's study, 0.48% of sessions to US manufacturing and industrial sites came from AI assistants and AI Mode between January and July 2026. Direct traffic was 56.65% and organic search 22.28%.
What is the difference between an AI mention and an AI citation?
A mention is a brand name appearing in the text of an AI answer, while a citation is a link the AI offers as a source. In Semrush's manufacturing data, only two domains, Vevor and Grainger, were in the top 15 for both.
If AI sends so little traffic, does AI visibility matter?
It may, because buyers who find a brand in an AI answer can arrive later through a direct visit or branded search that analytics does not attribute to AI. Semrush's data showed direct traffic share rising while organic search share declined, but the study does not prove that AI caused the shift.
Is this study independent?
No. It was produced by Semrush, which sells AI visibility tracking, and Semrush describes the combined analysis as directional rather than definitive.
Small AI referral numbers and real AI influence can both be true at once. If you want to measure where your B2B buyers are finding you, and whether AI answers are sending them, talk to Relevant Audience.
