Semrush today published a practical guide explaining how B2B marketers can grow their brand's visibility inside AI search tools by optimizing LinkedIn content - drawing on a dataset of 89,000 unique LinkedIn URLs cited by ChatGPT Search, Google AI Mode, and Perplexity.

The guide, authored by Margarita Loktionova, Product and Content Marketing Lead at Semrush, was published on June 9, 2026. It builds directly on research released in March 2026, which found that LinkedIn is the second most-cited domain in AI search responses, appearing in 11% of AI responses on average across the three platforms analyzed. That finding alone makes LinkedIn a meaningful surface for B2B brands attempting to influence what AI tools say about them.

The playbook lays out seven specific tactics. Each one is grounded in the dataset, not in general content advice. What distinguishes this guide from standard LinkedIn growth content is its focus on a single outcome - getting cited by AI systems - rather than engagement metrics or follower counts.

LinkedIn at number two across AI search platforms

The underlying data is striking. According to Semrush, the analysis covered 325,000 unique prompts submitted between January and February 2026. The 89,000 LinkedIn URLs drawn from those prompts came from ChatGPT Search, Google AI Mode, and Perplexity collectively. LinkedIn's 11% average citation rate places it above Wikipedia, YouTube, and every major news publisher in the dataset.

Citation behavior varies significantly by platform. According to Semrush, Perplexity cites Company Pages 59% of the time, while ChatGPT Search and Google AI Mode cite individual creators 59% of the time. That asymmetry matters for strategy. A brand that publishes only on its Company Page will reach Perplexity but will be largely absent from the citation pools of the two other major tools. Conversely, employee content fills the gap ChatGPT and Google AI Mode prefer - but only if those employees are producing consistent, well-structured posts.

The split between content formats is similarly nuanced. Across all three platforms, articles make up 50% to 66% of LinkedIn citations. Short posts account for 15% to 28%, depending on the model. Neither format dominates the other, and according to Semrush, they tend to surface for different kinds of prompts. Articles carry weight for in-depth answers, while shorter posts capture more specific, discrete questions.

Starting with customer questions - and using AI to validate them

The first recommendation in the playbook centers on identifying the questions buyers actually ask AI tools, then building LinkedIn content around those questions directly. According to Semrush, the best starting points are sales and customer success conversations, including demo recordings, support tickets, and win/loss notes - sources that capture the exact language buyers use.

The guide also points to review sites, social media comments, and community forums as places where buyer frustrations and questions surface organically. Once a list of potential content topics is assembled, the recommendation is to run those questions through an AI prompt research tool to verify which prompts are already generating AI answers in a given category, which brands get cited in those answers, and what related prompts might also be worth addressing.

This last step distinguishes the approach from conventional keyword research. In traditional SEO, a marketer optimizes for search volume. In the AI context described by Semrush, the question is which prompts a brand can own - meaning which questions, when asked of ChatGPT or Perplexity, currently return an answer that does not include that brand. Identifying those gaps gives content teams a focused brief rather than an open-ended mandate to "create more content."

The example Semrush uses to illustrate this principle is a LinkedIn article that was cited across 36 or more Google AI Mode prompts. The article answered a specific question about digital transformation planning. Its citation volume came not from virality or an engaged audience, but from its structure and the specificity of the question it addressed.

Employee advocacy as an AI visibility mechanism

The second tactic addresses employee advocacy programs - and frames them explicitly as an AI visibility strategy rather than a brand awareness play.

According to Semrush, the research shows that large language models prefer user-generated content because it signals first-hand experience. An employee who writes from direct professional experience produces something an AI system reads as more credible than polished company messaging, which can read as promotional.

The practical advice is specific. Semrush recommends identifying employees who have hands-on expertise and some interest in building a personal brand, rather than defaulting to executives or marketing staff. Engineers, product owners, and customer success managers are cited as often having more operationally specific knowledge that translates into citable content.

The guide also addresses the common failure mode of employee advocacy programs - the blank page problem. Telling employees to "share thought leadership" produces no output. According to Semrush, the solution is to provide hooks, templates, ghostwritten starting points the employee can edit, and tools that help them write in their own voice. Weekly time allocated for content creation and a small budget for tools such as LinkedIn Premium or design software are also mentioned as mechanisms to sustain participation over time.

For Semrush itself, according to the document, the practice involves leaders repurposing corporate research and data by adding their own commentary and analysis. This turns a single research asset into multiple pieces of content across different authors, each carrying a slightly different angle and reaching a different slice of the platform's audience.

Structure as citation infrastructure

The third tactic tackles the structural requirements of content that AI systems find easy to extract and cite. This is where the guide gets technically specific.

According to Semrush, AI tools frequently pull individual paragraphs out of a post when generating an answer. If a paragraph depends on the one before it to make sense, it will not stand alone as a citation unit. The recommendation is to write every paragraph so that it carries a complete idea independently - a requirement that pushes against many conventional writing styles where ideas develop progressively across a post.

Other structural recommendations include keeping paragraphs to one to three sentences, using numbered lists for parallel items such as steps or ranked comparisons, and spelling out acronyms on first use. The last point matters specifically in B2B contexts where specialist terminology - PAM, RAG, ABM, and similar abbreviations - may be misinterpreted by AI systems when context is missing.

One recommendation cuts against instincts toward brevity or brand subtlety. According to Semrush, vague references to products or categories get paraphrased into generic language when AI summarizes content. Writing "Semrush's AI Visibility Toolkit" rather than "our platform" keeps the brand name attached to the citation. This is a small formatting decision with a direct commercial consequence.

The guide cites a specific example - a LinkedIn article by John Shehata that ranked as one of the most-cited URLs in the entire Semrush study. That piece, a curated list of SEO newsletters, received only 31 likes and 12 comments. It was cited in 45 ChatGPT prompts. According to Semrush, its citation volume came from its structure - a clear opening, a numbered list, and sections that each answered a specific question independently - rather than from audience engagement. That disconnect between engagement metrics and AI citation performance is one of the study's central findings, and it has direct implications for how marketing teams measure content success.

The point-of-view requirement

Tactics four and five address what the guide calls differentiated positioning - taking a clear, specific stance on a topic and maintaining consistent terminology across all LinkedIn content.

The reasoning is data-based. According to Semrush, AI systems have access to thousands of sources on any given topic. When posts in a category say roughly the same thing, the algorithm has no particular reason to prefer one source over another. A clear, differentiated position backed by evidence - a specific number from a project, a customer result, a test outcome - gives the AI a reason to select that source over a generic alternative.

Semrush describes this in practical terms: frame the topic with a position that readers can either agree or disagree with, lead with the take immediately rather than building toward it, then support it with concrete proof. The guide also recommends providing a solution - explaining not just what is wrong with the status quo but what to do differently.

On terminology, the recommendation is to build a brand glossary and use it consistently across all authors, including executives and ghostwriters. According to Semrush, when the same terms and category names appear consistently across an account's content, AI systems build a stronger association between the brand and that category. Switching vocabulary weakens that association even if the underlying content is high quality.

The example used is Ina Nikolova, a cybersecurity professional cited across multiple AI engines for her content on Privileged Access Management. According to Semrush, she uses the same vocabulary - PAM, managed services, identity - across all her LinkedIn content, defines those terms explicitly in each piece, and has consequently built strong entity recognition in AI systems covering her category.

Series logic and citation clustering

The sixth tactic involves organizing content into topical series rather than publishing standalone pieces. According to Semrush, writing a central long-form article and then publishing a sequence of shorter posts derived from it builds what the guide calls topical authority - a sustained AI association between an author and a subject.

This approach treats a research report or detailed analysis as the beginning of a content program rather than a one-off publication. The example Semrush provides involves a coupon-related report - The Krazy Coupon Lady's 2025 State of Couponing report - which generated multiple LinkedIn posts including a video by a company executive, a post summarizing key findings, and employee content. Several of those posts ended up cited by AI tools.

The structural logic is that AI systems, when answering a question about a topic, look for consistent, recurring sources. A brand or individual that publishes once on a topic may generate a citation. One that publishes repeatedly, with linked angles and consistent terminology, becomes a recurring reference point. The series model is a practical mechanism for building that kind of sustained presence.

Frequency over virality

The final tactic addresses the relationship between engagement and citation, and explicitly separates the two metrics.

According to Semrush, approximately three-quarters of the cited authors in the study were frequent posters - producing five or more posts in the previous four weeks. Frequency and credibility outperformed reach. The John Shehata example recurs here: medium engagement, 45 ChatGPT citations. The point is that LinkedIn's algorithm rewards engagement, while AI search rewards clarity, structure, and consistency. Both matter, but virality is not required for either.

The guide recommends tracking AI citations alongside conventional engagement metrics and looking for patterns - specifically, which types of content generate citations versus reactions. This requires tools beyond LinkedIn's native analytics. Semrush positions its own AI Visibility Toolkit as one mechanism for this tracking, though the guide's structural recommendations apply independent of any particular measurement tool.

Why this matters for marketing professionals

The broader context for this playbook is a significant shift in how B2B buyers research vendors. According to Semrush, 95% of B2B buyers who use AI rely on it to research vendors and solutions. Out of more than 600 people surveyed for the study, 84% use AI search tools directly during the buying process.

That figure connects to a pattern that PPC Land has documented across 2025 and 2026 - the upstream influence problem. By the time a buyer clicks through to a website or requests a demo, their preferences may have already been shaped by AI-generated summaries comparing vendors, ranking options, and producing shortlists. That influence does not appear in last-touch attribution. It is, however, happening. Research published in May 2026 found that 44.8% of organisations have increased PR spending specifically because of AI-driven search, recognizing that AI recommendations cannot be bought the way paid search placements can.

The Semrush playbook arrives at a specific juncture in this discussion. LinkedIn's own research published in late 2025argued that brands need to build what it called "owned prominence" - visibility that accumulates through content and brand assets rather than through paid placements. The Semrush dataset now provides empirical support for one specific mechanism within that argument: structured LinkedIn content, published consistently by both brands and employees, is measurably influencing what AI systems say in response to B2B buyer queries.

The measurement challenge that PPC Land has covered extensively - that 81% of B2B marketing leaders consider AI visibility a blind spot in their organisation - makes the structural guidance in the Semrush playbook practically useful. Most marketing teams cannot directly observe whether they are being cited in AI responses. What they can control is the structure, consistency, and specificity of their content. Those are, according to the data behind this guide, the variables that determine citation outcomes.

For PPC Land readers focused on paid media, the implications extend beyond organic strategy. As documented in March 2026, the citation rate difference between platforms - Perplexity at 5.3%, Google AI Mode at 13.5%, ChatGPT Search at 14.3% - reflects distinct source preferences at each AI tool. Paid placements in AI interfaces, which Google began rolling out in AI Mode in May 2026, operate alongside organic citation logic. A brand that ranks well organically in AI citations may find that its paid activity in those same surfaces reaches a more primed audience - one already familiar with the brand through prior AI-generated answers.

Timeline

Summary

Who: Margarita Loktionova, Product and Content Marketing Lead at Semrush (Adobe), authored the playbook. Semrush, the Boston-based online visibility management platform currently being acquired by Adobe for $1.9 billion, published it. The intended audience is B2B marketers, content strategists, and SEO professionals managing brand presence across AI search platforms.

What: A seven-part operational guide explaining how to structure, publish, and organize LinkedIn content to increase the probability of being cited by AI search tools including ChatGPT Search, Google AI Mode, and Perplexity. The guide is based on an analysis of 89,000 LinkedIn URLs cited by those three platforms, drawn from 325,000 AI prompts. Key recommendations cover content structure, employee advocacy programs, consistent terminology, point-of-view differentiation, topical content series, and measuring AI citations separately from engagement metrics.

When: The guide was published on June 9, 2026. The underlying research was conducted between January and February 2026 and first published on March 10, 2026.

Where: The guide was published on LinkedIn's marketing blog. The research covered content cited across ChatGPT Search, Google AI Mode, and Perplexity globally, with a concentration in B2B categories including technology, business services, and finance.

Why: B2B buyers are increasingly using AI search tools to research vendors and make purchasing decisions before engaging with sales teams. According to Semrush, 95% of B2B buyers who use AI rely on it to research vendors, and 84% of those surveyed use AI search tools directly during the buying process. Because AI-generated recommendations influence decisions upstream of any click or form fill, brands that do not appear in AI responses are invisible at a critical stage of the buyer journey. The playbook gives content teams a concrete framework for addressing that gap through LinkedIn - a platform that the underlying data shows is the second most-cited domain in AI search globally.