← Back to curated
Afternoon BriefAI Search & Discovery

The Exact AI Search Taskforce LinkedIn Built After Losing 60% of Its B2B Traffic

LinkedIn's own content team lost 60% of B2B awareness traffic to AI Overviews, while rankings stayed put. Their internal fix: a cross-functional AI Search Taskforce. Here's the exact structure, and how to adapt it to your org this week.

Christian Lehman|
The Exact AI Search Taskforce LinkedIn Built After Losing 60% of Its B2B Traffic

LinkedIn's B2B content team didn't see it coming. Rankings were holding. Their SEO was clean. And then their non-brand awareness traffic dropped by up to 60%.

Nothing broke. Nothing was penalized. Google's AI Overviews simply ate the click before it could reach the page, across a subset of B2B topics, while keyword positions stayed exactly where they were.

That's the scenario most B2B marketing teams will face this year, and many are already inside it without knowing it. If your traffic is softening while your keyword positions hold, that's not a performance issue. That's a citation infrastructure problem. The two fixes don't overlap.

LinkedIn's internal team documented exactly what they did. Here's the model, stripped to what matters operationally.

Why Rankings Became a Lagging Indicator

The old equation: rank on page one, earn the click, convert the visit. That chain assumed search returned ten links and users clicked through to evaluate them.

A 2024 SparkToro and Similarweb study found that roughly 60% of Google searches in the US and EU already end without a click, and that number has continued rising since. AI Overviews accelerated it by surfacing an answer directly inside the results page, before the list ever loads.

The problem for B2B marketers: traffic from organic rankings and visibility inside AI-generated answers are now two distinct things. You can be ranked first and still be invisible to anyone whose query gets resolved by an AI summary. LinkedIn's team confirmed this explicitly, click-through rates softened even as rankings stayed stable.

Which means if you're measuring your organic program by traffic and clicks, you're watching the wrong signal.

What LinkedIn Actually Did

In early 2024, LinkedIn's B2B Organic Growth team started studying Google's Search Generative Experience. By early 2025, when SGE became AI Overviews at scale, the traffic impact was clear. Their response wasn't to publish more content or run a technical SEO audit. They built a cross-functional task force.

The LinkedIn AI Search Taskforce spanned eight functions: SEO, PR, editorial, product marketing, product, paid media, social, and brand. Their mandate wasn't to rank better. It was to show up in AI-generated answers, to be seen, mentioned, and synthesized before a user made a decision.

Three specific actions drove their early results:

  1. Correcting AI misinformation. They actively identified places where AI answers about LinkedIn were factually wrong or incomplete, then produced content to correct the record. This is reputation management for machines, not humans.

  2. Publishing content optimized for extractability. Not just well-written content, but content structured so that AI systems could pull a clean, citable snippet. Specific claims. Named data points. Attributed quotes. Information hierarchy that a language model can parse at a sentence level.

  3. Testing LinkedIn social content as an AI citation signal. Their hypothesis: posts on the LinkedIn platform itself could reinforce what AI systems surface about LinkedIn as a brand. Early tests showed meaningful lift in citation frequency.

Their early measurement: "triple-digit growth in LLM-driven traffic", still a small absolute number, but the directional signal matters. According to Semrush's three-month citation study, which analyzed over 100 million AI citations across ChatGPT, Google AI Mode, and Perplexity from July through October 2025, LinkedIn ranked among the top five most-cited domains on all three platforms. That structural advantage didn't happen by accident.

The Adapted Playbook for Operators

LinkedIn has a built-in moat, their platform content appears in AI answers partly because LLMs are trained on it. Most B2B brands don't have that. But the taskforce model is replicable at any org with the right structure and prioritization.

Step 1: Run a citation audit before touching your content.

Your first move isn't content production. It's intelligence gathering. Go into ChatGPT, Perplexity, and Google AI Overviews and run your ten most important buyer-intent queries. Track: Does your brand appear? What does it say? Is any of it wrong, incomplete, or missing context that a competitor is providing instead?

This takes two hours. Most teams skip it and go straight to publishing more. That's backwards.

Fixing what AI already says about you is higher leverage than publishing new content that AI hasn't indexed yet.

Step 2: Fix the extractability problem, not the word count.

AI systems don't need longer pages. They need pages that make specific claims in structured, parseable sentences. According to research from Content Marketing Institute and SAP's global SEO lead, LLM-referred visitors convert at twice the rate of organic traffic, but only when the content is structured clearly enough to be understood, cited, and acted on.

The fix isn't a content overhaul. It's a structural pass:

  • Add schema markup (Organization, Article, FAQPage)
  • Write a TL;DR or summary block at the top of key pages
  • Make every factual claim a standalone sentence, not embedded in a paragraph
  • Ensure named data points are attributed and linkable

Step 3: Assign citation-earning the same standing as rankings.

LinkedIn's task force worked because it crossed functional lines. SEO alone can't fix an AI citation gap. PR affects what third-party sources say about you. Editorial controls what your brand says about itself. Product and brand control what facts exist to be cited.

Most B2B teams run these functions in separate lanes. AI citation performance requires them in the same room with a shared mandate: what does an AI say about us, and how do we change it?

You don't need eight functions. Even a three-person group, SEO, content, and PR, running a monthly citation review is a structural upgrade over no coordination at all.

Step 4: Track AI penetration as its own metric.

The Semrush study showed that citation share can shift dramatically in weeks, as it did when ChatGPT changed how it sourced Reddit and Wikipedia in September 2025. Brands that weren't monitoring had no early warning. The measurement stack most teams are missing: Google Analytics 4 filters for AI referral sources, weekly manual prompt audits on core queries, and a citation baseline across the three platforms. None of this requires new tooling. It requires someone assigned to look at it.

The Visibility Gap You Can't See in Your Dashboard

LinkedIn's own marketing team, with more domain authority, more editorial output, and more platform advantages than most B2B brands, still lost 60% of awareness traffic to AI Overviews. The gap wasn't content quality. It was infrastructure for the new discovery layer.

Your brand likely has multiple citation gaps already forming. The question is whether you find them through a proactive audit or through a traffic report that arrives six months too late.

This is what Machine Relations describes as the operational shift B2B marketing teams are now inside, whether or not they've named it: the work of ensuring that AI systems surface, cite, and recommend your brand, not just rank your pages. LinkedIn's task force is one of the first large-scale documented examples of a B2B team building infrastructure specifically for that layer.

The prompt test is free. The audit takes two hours. The citation gap won't close on its own.

Related Reading

If you want to see exactly where your brand stands in AI-generated answers before your competitor does, the AT visibility audit shows you the specific queries where you're missing, what's being said instead, and where to fix it first.