Foursquare teases 'Ask' POI beta
Alex Kehr (ex‑Foursquare) signaled a beta 'Ask' endpoint for POI data validation, a product move that could tighten place‑data accuracy for developers and advertisers. (x.com)
Foursquare published an Ask API as a natural‑language place search endpoint on March 10, 2026. (foursquare.com) The Ask endpoint supports an optional context parameter that enables query‑specific personalization without storing user profiles, with examples like “dinner near me” plus context “allergic to gluten.” (foursquare.com) Every recommendation from Ask includes a transparent justification that cites user tips and relevant content, and the system uses an “agentic reasoning” multi‑step process for complex trade‑off queries. (foursquare.com) Foursquare reports its underlying Places data covers 100M+ POIs across 200+ countries and is backed by 16B+ check‑ins and 1B+ photos/tips, providing a large evidence base Ask can reference for validation and auditability. (foursquare.com) The Ask engine adapts execution strategy by query complexity: simple categorical searches complete in under 200ms while multi‑constraint queries finish in under 2 seconds, per Foursquare’s performance figures. (foursquare.com) Foursquare’s Places product combines human, software, and third‑party verification workflows for POI freshness and consistency, which Ask’s transparent signals and reasoning could surface into programmatic POI correction or flagging. (docs.foursquare.com) Developers can trial Foursquare’s Pro endpoints with test allowances—Foursquare advertises free testing for select Pro API endpoints (up to 10,000 calls) and pay‑as‑you‑go pricing for production usage. (foursquare.com)
