Google Maps Platform AI & Agents

AI & Agents. Location intelligence that agents can find, use and trust.

An assistant answering a question about peak traffic on a city road, beside the road on a satellite map
Grounded answersAn assistant answering a question about traffic on a road, with the road on the map.

AI models can’t guess whether your store is open today. That answer needs ground truth data. We connect agents to Google Maps Platform and to your own location services, so they answer from sources you trust.

The connections. How agents reach trusted location data.

AI assistants such as Claude, ChatGPT and Gemini connecting to your web services through MCP AI assistants such as Claude, ChatGPT and Gemini connecting to your web services through MCP

Model Context Protocol

The open standard for connecting AI to tools and data. Make your existing web services usable by Claude, ChatGPT, Gemini and others.

An assistant explaining traffic patterns with points of interest from Google Maps

Grounding with Google Maps

Answers anchored in Google Maps, with place detail and insights from reviews. Built for agents running on Gemini Enterprise.

An agent calling route and place search tools to plan stops on a drive

Grounding Lite

Google’s MCP server for places, routes and weather. It works with Claude, Gemini and other models.

An assistant answering a question about a neighbourhood, with an interactive map in the conversation

Google Maps AI Kit

Components that show an agent’s answer on an interactive map, inside the conversation.

Featured

Found by AI. Give agents the data they need to get your locations right.

Someone asks an assistant for the nearest branch. The answer needs accurate data.

Crawlable location pages and correct listings come first, and they feed AI answers on Google. An MCP endpoint then gives agents live detail, such as today’s opening hours.

A public location page with opening hours, fuels and facilities that AI assistants can read

What we build. Tools for agents, and agents for your teams.

Model Context Protocol logo

MCP endpoints

We turn your existing web services into tools agents can use. You choose what’s exposed, and every request is logged.

A skill file that tells an agent how to write a weekly corridor report

Skills

Teach agents to run your workflows, with no AI model to train.

A scheduled agent's run history, each run completed

Agent deployments

One specified task, run on a set cadence, with findings your team can act on.

Everyday needs counted within 5, 10 and 15 minutes by bike, on a map and in a table

Agents on your data

Plain-language questions across your BigQuery database and Google’s Insights data.

New from Google. Maps Imagery Grounding generates images and video of real places for marketing teams. Ask us what it could do for your next campaign.

Tell us which questions your customers ask. We’ll give agents the data to answer them well.

Other building blocks