What LinkedIn Learned From Publishing ~100 LinkedIn Articles

LinkedIn's senior digital marketing and AI search team, including Ina Mecklin, Cassie Dell, and Brooke Weller, recently shared what two years of internal testing has taught them about how brands show up in AI-powered search.

The session covered how LinkedIn itself had to rebuild its approach to content and visibility as SEO traffic began declining and LLMs became the primary discovery layer for many B2B buyers.

The central argument from the team was that AI search is not a separate discipline but a forcing function for integrated marketing. LLMs evaluate brand credibility across owned content, earned media, and social presence simultaneously, which means teams that have historically operated in silos need to coordinate more tightly than ever.

A few findings stood out. Earned media accounts for close to 10% of all LinkedIn's own citations in AI search results, making media relationships as important as technical optimization. On LinkedIn posts specifically, the first line of a post becomes its URL, which means hook-style opening lines actively work against discoverability. Articles in the 800 to 2,500-word range consistently outperform short posts for AI citations, and question-and-answer structure within posts improves the likelihood of LLM pickup.

On measurement, the team was candid. "The reality is there are still gaps," said Cassie Dell. "It is relatively a dark funnel, so try and do the best you can with the data pieces we have." Brooke Weller's framing on where to start was direct: "Do your own testing. Don't follow the hype. Figure out what works for your brand and bring that roadmap to your leadership."

Full story: Watch the event - Download the report

NEW AI LEARNING SERIES