With the widespread adoption of generative AI, companies are increasingly utilizing it for operational efficiency and decision-making support. However, the information that most current generative AIs can reference and utilize is centered on publicly available information and internal company documents. Integration with “geospatial data” and “people flow data,” which capture real-world conditions (“what is happening where right now”), remains insufficient.
Furthermore, technology trends are shifting from a single AI generating answers to a “multi-agent” model where multiple AI agents collaborate to execute tasks. Nevertheless, specialized data held by companies is often scattered across individual systems, and mechanisms allowing AI to seamlessly cross-navigate and utilize such data remain limited.
In addition to the geospatial data accumulated over many years, GeoTechnologies possesses multifaceted data including people flow, purchasing, and imagery. By providing this data in an AI-friendly format, GeoTechnologies aims to build a new social infrastructure in the AI era that supports decision-making for companies and local governments.

1. Autonomous Processing via Natural Language
The AI interprets requests made in natural language (everyday speech) and automatically coordinates and executes multiple steps, such as identifying target location coordinates, extracting data, conducting advanced analysis, and creating diagrams.

2. Support for the Open Standard “A2A Protocol”
GeoTechnologies has adopted the “A2A (Agent-to-Agent) protocol,” which supports collaboration between AI agents. This allows external generative AIs, orchestrators, and business systems to seamlessly delegate and coordinate tasks with GeoTechnologies’ AI agents using standardized procedures.

3. “Dependable” AI-Ready Design
GeoTechnologies built a design that allows enterprises to use the system with peace of mind in actual operations by providing the rationale behind analysis results and potential candidates, maintaining user context, and automatically recording operation logs.

Among these features, the major evolutionary step in this update is the support for the “A2A Protocol.” By standardizing integrations that were previously designed individually, tasks from data acquisition to deliverable creation can be executed as a unified workflow, enabling the automation of complex analytical operations. Users can leverage the location and people flow data owned by GeoTechnologies without changing the generative AI or business systems they are accustomed to using.
GeoTechAgent adopts a hybrid architecture combining MCP (Model Context Protocol), which connects “agents to tools and data,” and A2A, which connects “agents to agents.”

The “Purchase Analysis Agent” implemented in GeoTechAgent is an AI agent that combines receipt-derived purchase data with people flow data containing demographic attributes (such as age and gender) to clarify “who is buying what, where, and how much.”
Whereas specialized analysts previously conducted advanced consumer behavior analysis using multiple tools, users can now simply make requests in natural language. The AI automatically determines the optimal data and methods based on the request and outputs results by combining multiple analyses.
For example, entering a prompt such as “Output the beverage categories frequently purchased by women in their 30s in this commercial area, along with a market share comparison with competing brands” will automatically merge people flow and purchase data. It immediately generates and presents a report covering demographic purchasing trends to competitor shares.

< Sample Purchase Analysis Dashboard Interface >

• Sales & Purchase Trend Analysis: Visualizes sales trends by category and period, year-over-year comparisons, top-selling rankings, cross-purchase (basket) analysis, and trial-to-repeat conversion rates.
• Manufacturer & Brand Analysis: Examines sales summaries for specific manufacturers, regional purchase distributions, store channel configurations, and major product rankings.
• Competitor Brand Comparison: Quantitatively evaluates 1-on-1 comparisons between manufacturers, category market shares, and buyer segment overlaps.
• Customer Attributes & Persona Extraction: Derives demographic breakdowns (age, gender, etc.), heavy buyer profiles, and persona-specific purchasing characteristics.
• Integrated Visitation × Purchase Analysis: Merges trade area (people flow) data with purchase data to comprehensively track “what items were actually purchased by visitors to a specific area.”

Data Privacy Notice: The purchase data and people flow data used by each agent are obtained with user consent and statistically processed so that individuals cannot be identified. Analysis targeting individual behavior is not conducted. For details, please visit the introduction site.

Even for processes that take time to generate—such as insight reports and infographics—the task management technology of the A2A protocol visually displays progress in real time while generating outputs. Furthermore, because the rationale for analysis is clearly stated in the outputs, the results can be directly applied to practical marketing strategy development and product development decision-making.

GeoTechAgent also conducted proofs of concept for the following use cases utilizing map, people flow, and imagery data. In all cases, simply making a request in natural language allows the agent to automatically select the necessary data and analysis to derive optimal results.

GeoTechAgent Introduction Site:
https://english.geot.jp/products/enterprise_solution/geotechagent/

“We have successfully given form to the GeoTechAgent vision announced in April 2026 as a fully functional agent. By supporting the open standard A2A—in addition to MCP server connections for generative AI—we can now see a clear path for customers to directly call and utilize our map, people flow, and purchase data from their own generative AIs and business systems.
Going forward, we will build upon the insights gained from this PoC to sequentially expand supported data and features. We aim to contribute to the construction of a next-generation infrastructure that supports decision-making for enterprises and local governments by implementing autonomous AI that understands location, time, and context into society.”