Researchers have developed an Agentic Multi-Source Grounded system to improve query intent understanding in large marketplaces. This system grounds LLM inference in a catalog entity retrieval pipeline and an agentic web-search tool to handle ambiguous queries. It outputs an ordered multi-intent set that is then disambiguated using business policies, achieving 90.7% accuracy on DoorDash's search platform. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT This system offers a generalizable paradigm for grounding foundation models in proprietary context and real-time web knowledge to resolve ambiguous decision problems at scale.
RANK_REASON The cluster contains an academic paper detailing a new system for query intent understanding. [lever_c_demoted from research: ic=1 ai=1.0]