AI semantic search and ranking

Users search by needs and attributes, not only exact place names, so keyword search often misses intent.

- Improve place discovery. - Return ranked results with reasons. - Handle natural-language needs like quiet places to work.

People who search a map for places by their needs and attributes.

Search Query Dataset covering different search intents. POI Dataset with places, businesses, categories, locations, brands, and attributes. POI Metadata with ratings, popularity signals, descriptions, and tags. --- *Track: Mobility | Sponsor: Tasco | Slug: mobility:maps-semantic-ranking | Source: Agentic AI Build Week 2026 — Production snapshot 2026-08-11*

Query Input: user search query. Embedding Layer: generate query embeddings. Retrieval Layer: retrieve candidate results. Vector Database: store embeddings. Ranking Layer: AI-based ranking. Search API: return ranked results.