The news: LinkedIn’s AI-driven people search lets users type plain-language queries like “marketing leaders with AI experience” and instantly find matches—even if those exact words don’t appear on a profile, per TechRadar.

The upgrade, available to US Premium subscribers, makes LinkedIn far more context aware—and strengthens its role as a precision targeting engine at a time when its ad business keeps climbing.

The feature is a leap in semantic search, but the real innovation is the process behind it that shows how to scale generative AI (genAI) search across 1.3 billion members.

Smarter search for 1.3 billion users: Instead of relying on one monolithic AI model, LinkedIn’s engineers trained a small set of real search examples.

• Engineers used that data to teach a series of smaller, faster, specialized models. This halved the main model’s size with minimal accuracy loss, making it fast enough for billions of queries.
• They also rebuilt the infrastructure. While job search ran on traditional servers, the complexity of people search demanded a shift to powerful AI GPUs. 
• The search process was then split into a two-step “retrieve-and-rank” system: One model casts a wide net for potential results, and a second, precision model ranks them for relevance.

Zooming out: We forecast LinkedIn’s US ad revenues to grow from $4.86 billion this year to $5.67 billion in 2027. But growth in ad revenues is slowing, making AI-driven relevance and engagement an essential enhancement to keep advertisers invested. 

Why this matters for advertisers: Smarter search tightens the connection between user identity, knowledge, and engagement. For marketers on LinkedIn, the implications are significant and offer:

• Precision targeting. Search moves beyond job titles and companies to home in on nuanced expertise and inferred intent.
• Campaign efficiency. Higher relevance leads to improved engagement rates and stronger ROI for paid campaigns, content, and lead generation.
• Intent-driven discovery. Users can find communities and buyers based on what they know and what they’re seeking, not just what they’ve explicitly stated on their profile.