Delivered on July 21, 2026
A New Era Unleashed by TV Metadata x AI – M-Data's Vision for "Proper Data Utilization" and AI Agent Concept
Speakers:
Hiroyasu Taguchi, Executive Officer / Business Development Manager, Data Marketing Department, M-Data Co., Ltd.
Masaki Nishio, Account Executive ②
Affiliation and position information is as of the time of distribution
We welcomed M-Data Co., Ltd., which is celebrating its 20th anniversary, as our guest to discuss the potential of the TV metadata it provides and examples of its use. Based on highly accurate structured metadata of TV programs, commercials, products, people, companies, and tourist destinations, they will introduce their efforts to create new value in a wide range of fields, including marketing, advertising, tourism promotion, and local government policies. We also delved into the importance of data infrastructure in anticipation of the era of AI generation and AI agents, as well as the importance of creating data with quality guaranteed by human hands. This episode explores the future of how highly reliable data will support AI and business going forward.
＝＝＝＝
Definition of TV Metadata and the Four Pillars of Information Supported by M-Data
M-Data Corporation is a data supplier that celebrated its 20th anniversary in 2026, with the purpose of "making the world and business more interesting with data." The "TV Metadata" provided by the company is a highly structured database of television broadcast content, mainly composed of four categories.
The first is "TV Program Data," which subdivides programs down to the segment level and records details such as performers, keywords, and broadcast times. The second is "CM Data," which comprehensively covers all commercials for all companies and products that are broadcast approximately 4,000 times a day in the Kanto area alone. The third is "Product Data," which links not only the names and prices of products introduced in programs, but also JAN codes and purchase URLs, making it highly compatible with the retail industry. The fourth is "Spot Data," which databases facility information such as tourist destinations and restaurants, including latitude and longitude, which are essential for map services. These data are linked to proprietary master data on talent and companies, enabling a precise understanding of the overall picture of broadcasts.
A "Human-Centered" Data Generation Strategy Pursuing 100% Accuracy
Even in today's age of widespread AI generation, M-Data prioritizes data "accuracy" and maintains a human-centered data generation system.
At the data entry center in Mito City, Ibaraki Prefecture, approximately 130 operators watch television programs 24 hours a day, 365 days a year, and continuously input data in real time. The reason for not fully adopting AI is that the recognition accuracy of AI is still insufficient for "sudden changes in topic" and "simultaneous statements by multiple people (crosstalk)" seen in variety shows, etc. While extremely high accuracy is required for data for analysis, it has been determined that it is currently more efficient and cost-effective for humans to input the data from the beginning rather than for humans to correct "hallucination" (plausible lies) produced by AI. However, we are not completely excluding AI. We are actively utilizing it as a support tool in areas where it excels, such as input assistance, and are constantly seeking to balance accuracy and efficiency.
DX in the Retail Industry: AI POP and a 20% Increase in Sales
The application of television metadata is expanding significantly beyond traditional viewership analysis to the retail industry.
Efforts are underway to create "AI POP," which uses AI to learn television metadata and instantly reflect information featured on television on in-store price tags and POP displays. In a joint demonstration experiment with NTT West and others, a remarkable effect was confirmed: simply installing in-store POP displays utilizing television information resulted in a 20% increase in sales. Furthermore, a system for verifying the effectiveness of television exposure is being built to measure how much it actually contributes to driving traffic to stores by integrating television trend information with pedestrian flow data and beacon data. In addition, research is progressing on how to contribute to the efficiency of the entire distribution system, from predicting food trends to developing new product menus and optimizing order quantities based on demand forecasts.
Applications in the Financial Sector: Stock Price Prediction Signals and Value as "Alternative Data"
The "timeliness" and "social influence" of television as a medium are attracting attention as new indicators in the financial and investment fields.
Research using television exposure information as a leading indicator has shown that signals of stock price fluctuations can be detected with a high probability of 80% within three weeks of the start of a stock price increase. Furthermore, data from investment management using television metadata has shown returns five times that of the TOPIX (Tokyo Stock Price Index), proving that television can be an important factor in investment decisions. Data including the "instantaneous volatility" of specific companies or products being featured provides valuable information for investors. This is accelerating the movement to redefine the value of television media not merely as a result of public relations, but as new "alternative data" that influences economic trends.
Strategic Use of Television Exposure Information in Local Governments and Tourism Promotion
Local governments and tourism-related DMOs are also utilizing television metadata to visualize how the attractions of their regions are being reported.
By comparing and analyzing the amount of television exposure in one's own municipality with that of rival municipalities, it is being used to objectively evaluate tourism policies and formulate future promotional strategies. Beyond tourism policy, there is a growing trend to investigate what kind of news was reported in which areas during disasters, and to use this information to review administrative responses and information dissemination. As part of inbound tourism measures, efforts have begun to incorporate television news into multilingual tourism concierge services linked with digital signage. By measuring and predicting how many people actually visited and traveled to spots featured on television, it is possible to visualize the direct contribution to the local economy.
Next-Generation Outlook: Collaboration with Generative AI and the "AI Agent" Concept
M-Data has set seamless collaboration with AI services as its most important strategy for its outlook towards 2026.
As part of this, "TV Rank AI (beta version)," utilizing an MCP server compatible with Anthropique's generative AI "Claude," has been announced. This allows users to receive answers based on the latest broadcast data to questions such as "What are the hottest tourist spots recently featured on TV?" through a dialogue format with AI. By feeding 100% factual TV data to the AI, the aim is to overcome the AI's weakness of hallucination and deliver highly reliable information. By leveraging the power of AI, which can analyze vast amounts of data from multiple angles, it is expected that the potential of the data will be multiplied many times over by discovering "signs of a hit" and "breakthrough patterns" that humans might not have noticed. In the future, there are plans to provide services as "AI agents" specialized for various industries such as media, tourism, and finance.
Summary
The "TV metadata" provided by M-DATA Co., Ltd. has evolved beyond mere broadcast records into powerful "fact data" for solving modern business challenges. The fact that the "human-centered" data generation system, with 130 operators, supports "data reliability," which is of utmost importance in the AI era, is highly symbolic.
Its applications are diverse, including increasing sales in the retail industry, forecasting in the financial market, and developing tourism strategies for local governments, but what is particularly noteworthy is its integration with the latest generative AI. As information becomes increasingly digitized, M-Data's efforts to continuously structure reliable sources like public airwaves will overcome the uncertainties brought about by AI and become an indispensable infrastructure for implementing accurate and valuable information into society.