Google Maps grounding reaches general availability in Vertex AI

Google announced on September 26, 2025, that Grounding with Google Maps in Vertex AI has reached general availability. The feature enables developers to build generative AI applications connected to real-time information from Google Maps, according to product managers Greg Brosman and Lukas Bruderer.

The announcement addresses a fundamental challenge in generative AI development: ensuring model outputs remain factual and reliable when users depend on them for real-world decisions. Grounding refers to the process of enabling large language models to leverage verifiable information sources, either from Google or customer data.

The feature provides access to geospatial context from over 250 million businesses and places worldwide. This includes data from Google Maps' community of contributors who share local insights.

Real-time business information

Grounding with Google Maps offers several capabilities for developers building AI applications. The system provides users with timely details such as current restaurant hours or temporary closures, drawing from daily-updated Google Maps information.

The feature enables applications to answer subjective questions about places by incorporating context from Google Maps user reviews. Examples include queries such as "What's the vibe of this cafe?" or "Is this a good place for a quiet dinner?" This capability currently works for places in the United States and India.

Developers can combine Google Maps data with their own business data to accelerate business discovery. A real estate company, for instance, could create property summaries tailored to different client profiles. A family with young children would receive summaries highlighting nearby parks and schools, while a young professional might see information focused on local nightlife and restaurants.

The system allows combining Grounding on Google Search and Google Maps to provide comprehensive answers. When a user asks "can we bring our kids to tonight's concert in the park?", Google Search can determine who is performing and venue policies regarding children. Google Maps then provides information on kid-friendliness from user reviews and nearby suggestions. Gemini can dynamically select when to use one tool or both. A search for "best things to do in Barcelona" would trigger Grounding with Google Search, while "is the Sagrada Familia open now?" would use Grounding with Google Maps.

Industry applications

Several industries have shown early adoption patterns. In travel and tourism, the feature helps travelers find recommendations for hotels, restaurants, and attractions throughout the trip experience—from accommodation selection to itinerary planning and on-the-go local suggestions.

Peggy Roe, Executive Vice President and Chief Customer Officer at Marriott International, stated: "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."

Real estate companies can provide potential buyers with detailed neighborhood information, including custom listing summaries and suggestions for similar neighborhoods. Yotam Lemberger, Vice President of Engineering at Compass, explained: "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."

Device manufacturers are incorporating the feature into personal assistants to help users discover nearby attractions and answer neighborhood questions. Social media platforms are implementing it to help friends jointly research activities and dining options, including group exploration of areas, restaurant suggestions, and local nightlife information.

Technical implementation

The feature is now available everywhere with Gemini and includes a tier for easy experimentation offering 10,000 prompts with Gemini Pro. Developers can add the Google Maps tool to the Gemini API to enable grounding on Google Maps.

The implementation requires importing the necessary modules from the Google GenAI library. Developers create a client instance and generate content with a specified model, such as gemini-2.5-flash. The configuration includes tools with Google Maps enabled and retrieval configuration specifying latitude and longitude coordinates.

A sample query like "Where can I get the best espresso near me?" with coordinates for latitude 40.7128 and longitude -74.006 would return responses listing top-rated places to get espresso near the specified location.

Significance for marketing professionals

This development holds particular significance for the marketing community. Search advertising has long relied on location data to deliver relevant results, and the integration of real-time Maps data into AI applications creates new opportunities for local marketing strategies.

The ability to answer subjective questions about places based on user reviews represents a shift in how consumers might discover businesses. Rather than scrolling through review listings, users can ask nuanced questions and receive synthesized answers drawn from community feedback. This changes how businesses should think about their online reputation and review management.

For travel and hospitality marketers, the capability to provide personalized recommendations throughout the trip experience opens new engagement opportunities. Marriott's implementation suggests major travel brands are already exploring how to integrate this technology, potentially changing how consumers research and book accommodations.

Real estate marketing stands to benefit from the ability to generate customized property information based on buyer preferences. The technology can automatically tailor messaging to different audience segments without manual intervention, making personalized marketing more scalable.

The combination of Google Search and Google Maps grounding introduces complexity for search marketers. Queries that previously might have returned pure search results could now trigger Maps data, or both. Understanding when and how each grounding method activates becomes important for optimizing content and local business information.

The experimentation tier offering 10,000 prompts with Gemini Pro lowers barriers for agencies and smaller marketing teams to test applications before committing to larger implementations. This accessibility could accelerate adoption across the marketing industry.

As AI-generated responses become more common in search experiences, ensuring business information remains accurate and up-to-date across Google Maps becomes increasingly critical for local marketing strategies.

Timeline

• September 26, 2025: Google announces Grounding with Google Maps in Vertex AI reaches general availability
• September 26, 2025: Feature becomes available everywhere with Gemini, including 10,000-prompt experimentation tier
• Related: Google Search now has 2 trillion searches per year
• Related: Google launches Merchant Center Next
• Related: Google Ads introduces Hotel Ads Performance Max

Summary

Who: Google announced the general availability through product managers Greg Brosman and Lukas Bruderer. Marriott International and Compass are among the early adopters implementing the technology.

What: Grounding with Google Maps in Vertex AI, a feature enabling developers to build generative AI applications connected to real-time Google Maps data covering over 250 million businesses and places worldwide.

When: The announcement occurred on September 26, 2025, with immediate general availability.

Where: The feature is available everywhere with Gemini. The subjective question capability about places currently works for locations in the United States and India.

Why: The feature addresses the challenge of ensuring AI model outputs remain factual and reliable for real-world decisions by grounding large language models in verifiable information from Google Maps, including daily-updated business information and community-contributed reviews.