Artificial intelligence, meet location intelligence

AI is excellent at many things, but often struggles with spatial concepts. This session from BUILD with Mapbox 2025 explores the integration of artificial intelligence and location intelligence to create more intuitive and powerful AI agents. Get an overview of how Mapbox is reimagining the future of maps with AI agents. Mapbox equips AI agents and Large Language Models (LLMs) with spatial awareness and geospatial skills via the Mapbox MCP Server. For developers building with Mapbox, the Mapbox DevKit MCP Server for use with coding assistants makes development faster and more enjoyable. And the experimental Mapbox Location Agent illustrates the potential for conversational agents to upend the conventional design of interactive maps and location experiences. Highlights: -The Problem: Many people spend a significant portion of their day navigating the real world, often bouncing between multiple apps to find locations, plan stops, get directions, locate parking and more. Current AI solutions often miss crucial real-world considerations when it comes to understand geospatial context and physical navigation. -The Mapbox Solution: Mapbox bridges the gap by bringing location intelligence and geospatial skills to LLMs and agents. -The Mapbox MCP Server: A wrapper for Mapbox APIs that transforms any AI agent into a geospatially aware system. It allows agents to understand and respond to location-based queries (e.g., "nearby," "10-minute walk away"). -The Mapbox DevKit MCP Server: Designed to help developers build geospatial applications faster by augmenting AI coding assistants. It allows for quick setup and creation of custom maps and routes. -The Mapbox Location Agent: A demonstration of how Mapbox services can pair with an AI agent interface to create a conversational, map-integrated experience for tasks like finding hotels with specific views, planning custom walks with coffee and food stops, and even planning diving trips while avoiding hazards. -Future Vision of Conversational Maps: Mapbox envisions a future where maps are conversational, allowing users to interact with them through natural language, eliminating the need to jump between applications and providing a richer, shared visual and spatial context. -Industry Disruption: The speakers identify industries ripe for disruption, including store finders with inventory availability, personalized trip planning, navigation with complex constraints (e.g., avoiding highways), and outdoor recreation route planning. -Challenges: The speakers emphasized the importance of understanding real-world spatial relationships in AI, as well as keeping pace with the rapidly evolving AI technology. Links: Mapbox MCP server Github: https://github.com/mapbox/mcp-server?tab=readme-ov-file#mapbox-mcp-server Mapbox DevKit MCP server Github: https://github.com/mapbox/mcp-devkit-server https://github.com/mapbox/mcp-server/blob/main/docs/hosted-mcp-guide.md Request Access to the Mapbox Location Agent (private preview): https://www.mapbox.com/forms/location-agent