Connecting Generative AI with Real-World Foot Traffic Data. Location AI Platform® Starts External Provision via API/MCP

Location AI Inc. (Headquarters: Shibuya-ku, Tokyo, Representative Director: Kazuya Ono, hereinafter "Location AI"), a company that develops AI analysis technology for location-based big data, has started providing the "LAP API (Beta version)" today, which allows the functions and data of the cloud-based foot traffic data utilization platform "Location AI Platform® (hereinafter, LAP)" to be connected with external generative AI systems. Furthermore, in July 2026, they will start providing in the "MCP (Model Context Protocol)" format, a standard connection specification for generative AI.

Since its inception, Location AI has consistently engaged in the research and development of technology to statistically, estimate, and numerically process "real-world data" such as location information using AI. The core of this is the proprietary analysis engine "Location Engine™," and the platform "LAP" for visualizing and analyzing results, both of which are entirely developed in-house.

With the provision of API/MCP this time, it becomes possible to connect and supply "real-world foot traffic data that changes moment by moment," which each model cannot possess as learning data, to various models of generative AI that companies introduce and operate in-house, such as ChatGPT, Claude, and Gemini.

The reason why foot traffic can be provided as machine-readable data via API, without human aggregation or interpretation, is that the process of converting real-world foot traffic into "data" that is completely structured by AI technology has already been completed as a technology in Location Engine™.

Users will be able to extract insights that previously required specialized analysis work, simply through natural language questions and dialogues, by combining the advanced inference and analysis capabilities of generative AI with real-world foot traffic data.

■ Background: 80-90% of economic activity occurs in the "real world"

Even as e-commerce advances, 80-90% of consumer behavior and economic activity, such as retail, dining, services, and face-to-face business, still occur in the real world (the e-commerce penetration rate for retail is approximately 10%, and for dining, over 97% is based on in-store visits*). Therefore, most of the business activities of companies and public organizations are based on the real-world context of "where people are actually located and what they are doing." Information including foot traffic data is an extremely important element for management decision-making and business judgment.

Meanwhile, as the business use of generative AI rapidly expands, generative AI's learning data does not include real-world foot traffic data that changes moment by moment. No matter how smart AI becomes, it cannot support high-precision business decisions without the context of the real world. Location AI, which has specialized in AI analysis of real-world data, provides "grounding" that connects AI and the real world, based on data assets of approximately 93 million device IDs domestically, approximately 4.2 billion IDs globally, and approximately 3 trillion records accumulated over 8 years, along with proprietary AI analysis technology.

This API/MCP provision is an initiative to open up foot traffic data as "fundamental data" for the generative AI era, from the world where it was conducted as "LAP AI Assistant" within the LAP user interface, to external generative AI systems. The more the use of generative AI expands, the more the demand for real-world data, which forms the basis of its judgment, will also expand. LAP is designed as a platform whose scope of use will expand along with the spread of generative AI.

*Source: Ministry of Economy, Trade and Industry EC Market Survey, Yano Economic Research Institute, etc. (2024)

■ Past Initiatives: From AI Utilization within LAP to Collaboration with External Generative AI

So far, the utilization of foot traffic data has typically been limited to data collection and visualization, or relying on human report creation for the final analysis process. Location AI has consistently pursued the automation of the entire process with AI technology, and since the implementation of "LAP AI Chat with ChatGPT" in LAP in 2024, it has continuously strengthened its foot traffic data analysis and utilization functions, starting the provision of "LAP AI Assistant" in 2025, where generative AI directly analyzes foot traffic data and provides analysis support in a dialogue format.

While receiving support from many users, the need for "directly integrating foot traffic data into the systems or generative AI environments operated in-house and utilizing dynamic analysis for business and operational decisions" has increased, and we have received many inquiries, including from existing LAP users.

■ LAP API (Beta version) Function List (as of June 2026)

Category

API Function

Overview

POI Management (Analysis Points)

Registration, Update, Deletion, List Retrieval

GeoJSON compatible, batch registration possible

Folder Management

Creation, Update, Deletion, List Retrieval

Organization and hierarchical management of POIs

Analysis Groups

Batch Creation, Deletion, List Retrieval

Automatic generation of condition combinations, dry-run preview support

Visit Conditions

List Retrieval

Retrieve visit condition IDs associated with groups

Foot Traffic Reports

Daily, Hourly Retrieval

Number of visitors and stayers by gender and age

Map Data

Visit Rate, Potential, Ranking

JSON retrieval by mesh unit

CSV Download

4 types of CSV downloads

Visit Rate / Potential / Basic Trade Area / Ranking

*Other foot traffic data to be provided in the future: Residential area data by town and block / Concurrent usage rate / HotPlace / Other analysis data

■ MCP, which directly connects generative AI and "real-world foot traffic data," will also be provided starting in July

In July 2026, we plan to publish a secure collaboration function compatible with the new standard "MCP (Model Context Protocol)" for standardly connecting generative AI LLM with external systems. With MCP compatibility, the collaboration between self-built generative AI environments and foot traffic data will be further simplified, allowing corporate DX promotion personnel and service developers to easily and seamlessly incorporate advanced foot traffic intelligence into their own generative AI systems.

By being compatible with the industry standard MCP, any company or organization that introduces generative AI in the future can become a user of LAP's foot traffic data without any special development.

For example, simply asking your own generative AI in natural language like the following can obtain answers based on real-world data as follows.

"Summarize the number of visitors to the flagship store in Shibuya this week compared to last month."

"Compare the trade area potential of three new store candidate locations, and show the recommended ranking and reasons."

"Propose improvement measures for promotional strategies based on changes in the gender and age composition of visitors."

■ About the Use of "LAP API" Beta Version

The LAP API Beta version trial is currently available to customers who have contracted with LAP, providing an environment where API keys are issued for use. After the MCP is provided, it is also planned to be provided to companies other than LAP users. If you are interested, please contact the inquiry window.

■ About Location AI Inc.

Location information big data AI analysis technology research and development company

Keyword: