LinkedIn Stopped Promoting Articles. AI Never Got the Memo. Here's the 3-Channel Audit That Fixes Your Citation Rate.
A MarTech study of 1,000+ AI prompts found YouTube drives citations for 75% of B2B brands and LinkedIn articles for 37% - while Wikipedia and Reddit barely register. Here's the exact 3-channel audit to realign your team's efforts.
Key Takeaways:
- YouTube appears in the top 25 cited domains for 75% of B2B brands getting AI citations - yet most B2B teams have no YouTube citation strategy
- LinkedIn articles (which LinkedIn itself stopped promoting) are in the top 25 cited sources for 37% of B2B brands in AI answers
- Wikipedia and Reddit - channels teams spend real budget on - barely register: 60% and 83% of B2B brands respectively have zero presence in those citation pools
- Owned media is cited more than 2x as often as earned media across B2B categories
- Only 21% of B2B brands appear in more than 25% of AI answers to their most important queries
Most B2B marketing teams have a LinkedIn Articles problem. They wrote a few posts in 2021, watched the platform bury them in the feed, and killed the format entirely. No impressions, no engagement, no point.
That decision probably tanked their AI citation rate - and they have no idea.
MarTech published one of the most data-grounded B2B AI visibility studies yet: 1,000+ AI prompts tested across 29 brands using ChatGPT, Perplexity, Grok, and Google Gemini. The study was conducted using Brandi, a GEO measurement tool, and tracked which specific domains and content types were driving AI citations for each brand.
The results contradict most of what marketing teams are doing right now.
Here's what the data actually shows - and the 3-step audit to run this week to close your gaps.
The Platform Data Nobody Shares in Full
The headline number from the MarTech study is damning: only 21% of B2B brands appeared in more than 25% of AI answers to the queries that mattered most to their buyers. One-third appeared in fewer than 5% of AI answers.
That part gets quoted. What doesn't: which channels are driving citations for the brands that are showing up.
| Platform | % of B2B Brands with Platform in Top 25 Cited Domains |
|---|---|
| YouTube | 75% |
| LinkedIn Articles | 37% |
| Wikipedia | 40% (60% NOT in top 25) |
| 17% (83% NOT in top 25) |
Source: MarTech / Brandi, 1,000+ prompts, 29 B2B brands, 4 AI engines
This is not a niche dataset. This is 29 B2B tech companies - electronic components, semiconductors, industrial automation, SaaS, logistics, and engineering - with a mix of enterprise and midsize brands. These numbers reflect B2B reality, not consumer or startup dynamics.
The Two Channels You're Probably Wasting Time On
Wikipedia
The conventional wisdom for AI visibility is: get on Wikipedia, build an infobox, keep your entry current. For consumer brands, this is directionally correct. For B2B technology companies, the MarTech data says otherwise.
For 60% of B2B brands studied, Wikipedia didn't appear in the top 25 cited domains for any AI answer relevant to their category. For 30% of brands, Wikipedia wasn't cited at all in any AI response. This doesn't mean ignore Wikipedia entirely - brand definition queries may still pull from it - but treating it as a core AI citation lever for B2B is not supported by the data.
Reddit shows up in AI citation pools for 83% of B2B brands as a non-factor. ChatGPT has commercial access to Reddit, but for B2B tech categories, that access isn't translating into brand citations the way it does for consumer products or software with high community engagement.
The exception: if your product genuinely has product-category discussion on Reddit - real users comparing you to alternatives - that signal can matter for ChatGPT specifically. But manufacturing Reddit presence as a B2B citation strategy is almost certainly a misallocation of time.
The 3-Channel Audit (Run This Week)
Step 1: YouTube Gap Check (30 minutes)
Open ChatGPT, Perplexity, Grok, and Google AI Mode in four tabs.
Run the 3–5 queries your ideal buyer uses at the top of their research cycle. Examples for a B2B SaaS brand: "What's the best [category] platform for mid-market teams?" or "How does [category] solve [specific problem]?"
When a YouTube video is cited, note the channel and format. When no YouTube appears for your category queries, that's your gap.
The question you're answering: does your brand have YouTube content an AI would pull when a buyer researches your category?
When a generative AI summary appears in search results, users click traditional result links only about 8% of the time - which means being cited in the AI response itself is increasingly the only reliable surface area. A cited YouTube video is the difference between being in the answer and being invisible while your buyer forms their shortlist.
The fix isn't a production overhaul. It's 2–3 structured videos that directly answer your buyers' most frequent research questions: "How does [product] handle [specific use case]?", "What does [category] implementation actually look like?", "[Your product] vs. [competitor]: an honest comparison." Publish to YouTube with descriptive, specific titles. Not LinkedIn video. Not Vimeo. YouTube is the platform AI engines pull from at scale for B2B content.
Step 2: LinkedIn Article Sprint (4 weeks)
LinkedIn buried Articles as a promoted format in 2022. The platform doesn't push them in the feed anymore. If you're measuring LinkedIn Articles by organic reach, they're effectively dead.
But that's the wrong metric now.
LinkedIn Articles are Google-indexed, carry high domain authority, and have high crawlability for AI training pipelines and retrieval systems. The MarTech study's finding - that 37% of B2B brands in AI citation pools have LinkedIn articles in their top 25 cited sources - includes articles published years ago. The AI doesn't care that LinkedIn stopped amplifying them.
The four-week sprint: publish one LinkedIn article per week, each targeting a single buyer research question at the decision stage:
- "How should a [role] evaluate [category] vendors?"
- "What does [category] implementation actually look like for a [company size] team?"
- "What are the hidden costs of [category] that don't show up in demos?"
Structure each article with a clear question in the title, a direct answer in the first paragraph, and specific sub-points with headers. No thought leadership. No company updates. Decision-stage questions get cited. Brand commentary doesn't.
Step 3: Owned Media Architecture Audit (1 hour)
The MarTech study's most surprising finding: owned media is cited more than twice as often as earned media across B2B technology brands. Most AI citation advice focuses on earned coverage and third-party mentions. The data says your website is doing more work - or failing more quietly - than almost anything else.
Search Engine Land's 90-day AI visibility framework reinforces this: AI systems organize information around entities and questions, not just keywords. Pages that answer specific buyer questions - structured with clear headers, direct first-sentence answers, and specific examples - are the ones that get pulled.
Run this audit on your core pages:
Do your category and use-case pages lead with a direct answer to a buyer question? Not "here's what we do" - but "here's how to solve [specific problem], here's what good looks like, here's where companies get it wrong." AI engines retrieve answers. They skip product descriptions.
Do your FAQ and comparison pages have structured, extractable responses? Dense paragraphs are harder for AI systems to pull and attribute accurately. Headers, short direct answers, and specific data points are more retrievable.
Are you publishing content on partner and channel sites as well as your own? The competition for B2B AI brand visibility has moved off the SERP - and partner ecosystem content is one of the most underutilized signals. If you have integration partners, resellers, or trade associations where you can contribute structured content, those are high-leverage citation opportunities.
Why This Needs to Run Before Q2
Forrester's 2026 B2B predictions note that as B2B buyers increasingly use AI tools to gather faster insights, marketing and sales teams face mounting pressure to ensure their brand narrative is present - and accurate - in those AI responses. Nineteen percent of buyers using AI research tools already report feeling less confident in purchasing decisions because of inaccurate or unreliable information. Brands that aren't present in AI answers don't just lose visibility. They create a vacuum that competitors fill.
According to IDC, AI-mediated buying journeys are increasingly filtering vendors for relevance and trustworthiness before buyers ever reach sales. The research phase your team used to influence is happening in AI conversations - and the brands that show up aren't necessarily the biggest. They're the ones with content structured to be retrieved in the channels AI engines trust.
This is the execution layer of Machine Relations: not just building brand presence, but architecting that presence across the specific channels - YouTube, LinkedIn Articles, owned media - where AI engines actually look when a B2B buyer asks about your category.
The channel mix is different from what most teams expect. The MarTech data makes clear where to focus. If your team has been writing off LinkedIn Articles and treating YouTube as a top-of-funnel awareness play, this week is a good time to reassess.
Run the audit. Run the category queries in ChatGPT and Perplexity. See whether YouTube or LinkedIn content is driving citations for your category - and whether any of it is yours.
If you want a structured view of where your brand currently stands across AI engines by channel, the visibility audit maps your citation gaps so you're not guessing which levers to pull first.
Related Reading
- Machine Relations for MarTech Companies: How to Win AI Engine Authority in a 15,000-Tool Landscape
- AI Visibility for Consumer Brands: The 2026 Earned Media Playbook
FAQ
Why does LinkedIn Articles still work for AI citations if LinkedIn stopped promoting them? LinkedIn Articles are Google-indexed and carry the authority of the LinkedIn domain (one of the web's highest-DA properties). AI training pipelines and retrieval systems crawl based on domain authority and content structure - not on whether LinkedIn chose to amplify a post in its feed. An article published in 2022 with clear structure and specific answers to buyer questions can still be retrieved in 2026 AI responses.
What makes YouTube content get cited by AI engines vs just watched? AI engines pull YouTube when the video title and description directly match the query - not when the video is good or popular. A well-titled explainer video answering a specific buyer question ("How does [category] handle [use case]?") is more likely to be cited than a polished brand video with a vague title. Treat YouTube titles as query answers, not attention hooks.
Does this mean I should stop investing in earned media? No - but reframe it. Earned media's primary value for AI citation is authoritative coverage in industry trade publications, not aggregate media impressions. The MarTech study found a small number of niche trade publications drove most of the earned citation volume, often titles PR teams had dismissed as low-reach. The owned-to-earned ratio (2:1 in favor of owned) means don't trade off owned for earned - run both, but allocate accordingly.