If your LinkedIn reach has fallen off a cliff this year, you are not imagining it, and it is not a temporary dip. On 12 March 2026, LinkedIn’s engineering team confirmed the platform had replaced its entire feed ranking system with a single, unified AI model built on large language models, a change the industry has largely nicknamed 360Brew after the internal research behind it. 

According to the announcement covered by Search Engine Land, the new system uses AI generated embeddings and GPU powered ranking to understand what a post is actually about and connect it to a member’s professional interests, rather than relying on the patchwork of separate systems LinkedIn used before.

For anyone running B2B marketing or outreach on the platform, this is not a minor tweak worth a paragraph in a monthly roundup. It is a genuine rebuild of how visibility works, and it changes what actually generates leads.

What Changed Under the Surface

Before this update, LinkedIn’s feed pulled candidate posts from several separate systems, network activity, trending content, collaborative filtering and topic tagging, then scored each one largely in isolation. That structure rewarded volume and network size. Post often enough to a large enough connection base, and reach followed reasonably predictably.

The new system works differently in a way that matters for anyone doing B2B outreach. Posts and members are now represented as points in a shared semantic space, so the platform can connect related ideas even when they are phrased completely differently. 

A post engaging with supply chain resilience, for example, can now surface content about logistics software or manufacturing automation, because the model understands the underlying topic rather than matching keywords. Distribution has shifted from being driven mainly by who you are connected to, toward being driven by what you are demonstrably knowledgeable about.

What Got Rewarded, and What Got Quietly Punished

Dwell time and saves now outweigh likes. The platform is measuring whether people actually read a post, not just react to it. A handful of substantive comments and saves from the right audience now outperforms hundreds of quick reactions from people outside your target market.

Topical consistency matters more than posting frequency. Profiles that jump between unrelated subjects confuse the model’s understanding of what that person or brand is actually an authority on, which suppresses reach even when individual posts are strong.

External links are quietly deprioritised. Because LinkedIn wants attention to stay on the platform, posts that push traffic off-site tend to be distributed less aggressively than posts that deliver the value directly in the feed.

Engagement pods and generic AI text are actively detected and penalised. Coordinated fake engagement and obviously templated AI generated posts are now identified with high accuracy and suppressed, closing off a shortcut a lot of B2B accounts leaned on through 2024 and 2025.

The net effect: reach has genuinely fallen across the board this year for accounts still running the old playbook, while accounts that adapted to reward genuine topical authority and real engagement are, by most independent analysis, holding steady or growing.

Why This Is a Bigger Deal for Outreach Than for Content

It is tempting to file this under “content marketing news” and move on if your focus is outbound lead generation rather than organic posting. That would be a mistake. The old model of LinkedIn outreach, cold connection requests followed by an immediate pitch, was already losing effectiveness against inbox fatigue. 

A ranking system that now actively rewards demonstrated topical relevance makes the case for a different approach even stronger: content and outreach working together, where a prospect has already seen genuinely relevant posts from your profile before an outreach message ever lands in their inbox.

That is a coordination problem as much as a content problem, and it is exactly why we run LinkedIn outreach as a connected discipline rather than a bolt-on to whatever content happens to get published that week. An outbound message to a cold prospect who has never seen your name lands very differently than one to a prospect who has already engaged with a post that demonstrated real expertise on a problem they actually have.

The Same Shift Is Happening Across Search, Not Just LinkedIn

If this pattern sounds familiar, it should. Google’s AI Overviews and tools like ChatGPT and Perplexity are going through a strikingly similar transition, moving from matching keywords toward understanding topics and citing whoever demonstrates real expertise on them. 

The shift towards GEO and the underlying logic is close to identical to what LinkedIn just rolled out: semantic, AI driven systems are replacing simpler signal counting across nearly every platform businesses rely on to be found.

That is not a coincidence, and it is not something a business can address by making one isolated fix to its LinkedIn strategy while leaving its website content on the old playbook. It is the same shift showing up in two different channels at once.

What to Actually Do About It

• Pick a narrow, consistent topic lane for your company page and key individual profiles, and stay in it. Authority in one clear area now travels further than broad, scattered posting.
• Write for dwell time. Longer, specific, evidence based posts that genuinely reward reading beat short hot takes and generic AI-generated filler.
• Move external links out of the opening lines, or avoid them where the value can be delivered natively instead.
• Sequence outreach after engagement, not before it. A message that references a prospect’s own comment or a post they clearly engaged with converts at a different rate than a cold pitch to a stranger.
• Treat pods and engagement shortcuts as a genuine risk, not a growth hack. Detection has improved enough that the downside now outweighs any short term reach boost.

The Bottom Line

LinkedIn did not tweak its algorithm this year. It replaced the engine underneath it, and the businesses adapting fastest are the ones treating content and outreach as one coordinated motion rather than two separate workstreams run by two separate people with two separate goals.

Want a straight answer on whether your current LinkedIn approach still works under the new system? Book a free consultation and we will show you exactly where the gaps are.