LinkedIn adds AI-powered search to help users find people (techcrunch.com)

LinkedIn is rolling out an AI-powered people search that lets users find professionals via natural-language queries — for example, “Find me investors in the healthcare sector with FDA experience” or “Who in my network can help me understand wireless networks.” The feature extends a U.S. job-search tool LinkedIn launched earlier and will initially be available to premium members in the U.S., with broader rollout planned. The search bar will prompt with “I’m looking for…,” and early tests show people using it to find job leads, business partners, and career mentors more quickly than with traditional keyword-and-filter searches. Technically, this is a shift from lexical, filter-driven retrieval to semantic, AI-driven matching (likely embedding/LLM-backed), designed to surface relevant people even when users don’t know exact titles or keywords. That improves discoverability but brings framing sensitivity and accuracy issues — results can vary depending on phrasing (“YC” vs “Y Combinator”) and may surface noisy signals like badges. The move puts LinkedIn squarely in the broader wave of platforms adding conversational/AI search and raises questions about platform data access and policies for browser/agent usage. LinkedIn says it’s iterating on understanding and ranking to reduce mismatches as it scales.