We are excited to announce Grounding with Google Maps in Vertex AI is now Generally Available (GA). This feature helps developers build generative AI applications that are connected to real-world, up-to-date information from Google Maps.
Why this matters: A key challenge for developers when building with generative AI is making sure model outputs are factual and reliable, especially when people rely on that information for real-world decisions. „Grounding” is the process of enabling a large language model (LLM) to leverage verifiable sources of information from Google or a customer’s own data.
What this unlocks: Now, developers can build AI applications grounded in geospatial context from over 250 million businesses and places worldwide with Grounding with Google Maps, including its active community of contributors sharing local insights.
Let’s take a closer look.
Better, more personal results
Grounding with Google Maps helps developers build factual and reliable generative AI outputs in multiple ways:
- Give users timely details, such as a restaurant’s current hours or if a shop is temporarily closed, by giving AI applications access to daily-updated Google Maps information.
- Make it easy for your users to ask subjective questions about places, with answers that include context from Google Maps User reviews such as „What’s the vibe of this cafe?” or „Is this a good place for a quiet dinner?” *
- Make business discovery faster by combining Google Maps data with your business data. For example, a real estate company can create property summaries for different client profiles, so that a family with young children gets a summary highlighting nearby parks and schools. For a young professional, it might focus on local nightlife and restaurants.
- Give users the full picture on their queries by combining Grounding on Google Search and Google Maps. Now, when a user asks, “can we bring our kids to tonight’s concert in the park?”, Google Search can look up who is performing and whether the venue has any policies on children, and Google Maps can provide relevant information on kid-friendliness including insights derived from user reviews and suggestions for what’s nearby. And by enabling both tools, Gemini can dynamically select when to use one of them or both. For example, searching for the “best things to do in Barcelona” would use Grounding with Google Search, and a subsequent query on “is the Sagrada Familia open now?” would use Grounding with Google Maps.
*This capability is currently available for places in the United States and India.
Useful across industries
Grounding with Google Maps can be used in many industries. Some places where we are seeing early adoption in particular include:
- Travel and tourism: Help travelers with recommendations for hotels, restaurants, and attractions. This includes across the trip experience, from deciding where to stay, itineraries of things to do, and local suggestions when on-the-go.
“Delivering authentic and high-quality service is all about the details. Marriott Bonvoy aims to provide consumers access to the best experiences when they travel and more connected visual mapping capabilities along with integrated local information. This deepens our engagement with guests by transforming how they discover, plan, and experience the world with us.”
Peggy Roe, Executive Vice President and Chief Customer Officer, Marriott International
- Real estate: Give potential buyers detailed information about neighborhoods, including custom listing summaries and suggestions on similar neighborhoods.
“The future of real estate is an experience that includes GenAI to improve the buying and selling journey. We are partnering with Google to bring this to life using Grounding with Google Maps in Vertex AI and other innovative AI capabilities. This will enable us to offer more personalized location insights and empower Compass agents to provide the best in class buying and selling experience to their clients.”
Yotam Lemberger, Vice President of Engineering, Compass
- Devices: personal assistants help users discover nearby attractions and answer questions about a neighborhood.
- Social media: Help friends jointly research what to do and eat. This includes exploring areas as a group, suggestions on nearby restaurants, or helping users learn more about the local nightlife.
Get started today
Grounding with Google Maps in Vertex AI is now available everywhere with Gemini and comes with a tier for easy experimentation (10K prompts with Gemini Pro). To learn more and get started, visit our documentation, check out our interactive demo, and visit the Vertex AI Studio in the Google Cloud Console. Simply add the Google Maps tool to the Gemini API to ground on Google Maps:
from google import genai
from google.genai.types import GoogleMaps,Tool, GenerateContentConfig client = genai.Client() response = client.models.generate_content( model="gemini-2.5-flash", contents="Where can I get the best espresso near me?", config=GenerateContentConfig( tools=[Tool(google_maps=GoogleMaps())], # Use Google Maps Tool tool_config=types.ToolConfig( retrieval_config = types.RetrievalConfig( lat_lng = types.LatLng( latitude=40.7128, longitude=-74.006 ))), ),
) print(response.text)
# Example response: 'Here are some of the top-rated places to get espresso near you: ...'Python
Add comment