ART-TRA Co., Ltd.: The Potential of the "Knowledge Mapping" Patent Technology and Card-Style UI for Search-Free, Intuitive Information Discovery

You Can\'t Find What You Don\'t Know Without a "Proper Query"

Hello, everyone. My name is Reo Mizuno, CEO of ART-TRA Co., Ltd. Thank you for having me today.

Under our mission, "Create New Era with Technology & Art," we are working on reconstructing the touchpoints of information. Specifically, we provide a "discovery mechanism" that allows users to intuitively encounter new information and knowledge without having to search or summarize.

Today, with Google search and generative AIs like ChatGPT, instantly searching and summarizing information has become the default. However, there is a massive, often overlooked problem in these conventional systems: you cannot get a proper answer unless you input the "proper query." In other words, unless users have a certain level of prior knowledge or awareness, it is extremely difficult to look up something they "do not know." We believe that this "step of creating the query" itself is the biggest hurdle for users rather than returning the answer.

Furthermore, most corporate websites today are designed with a "tree structure (hierarchical)." While this structure is easy for creators to organize, it hinders "exploration" for users. For example, a student visiting a recruiting site to get an image of a job might get a more realistic concept by reading past press releases or award histories in different categories rather than cycling through the designated "employee introduction" page. However, with the current segmented tree structure, users cannot smoothly navigate to useful information in other hierarchies.

The same applies to purchasing behavior; most people are unaware of their latent desires until they actually see and touch things. Yet, the standard online approach is limited to broad persona targeting like "males in their 20s" or recommendations heavily biased toward past search history. We want to recreate a spacious information discovery experience online—similar to browsing through books stacked on tables at a physical bookstore, where you happen to find something interesting.

"Knowledge Mapping" Patent Technology and Card-Style UI to Re-Link Information Assets

The solution we provide consists of three main features.

First is "Knowledge Mapping," our patented technology that uses AI to re-link a company's past information assets, such as blogs and news, into a relationship graph.

Second is our unique algorithm that naturally generates related information that users might want to see next with a touch of surprise.

Third is the "Card-Style UI" that allows users to intuitively explore information by flipping cards without searching. We also hold a patent for this unique interface.

Implementing this system transforms corporate web marketing. Conventional SEO and content creation were one-off, flow-type strategies where companies had to "constantly publish new information" to attract users. In contrast, our system can connect past articles with current ones as layers—for instance, showing how a project started four years ago has developed today. This allows companies to effectively convey their "story" and history, breathing new life into vast archives of past articles.

PoC Progress and Expansion Strategy from B2B to B2C

Leveraging our intellectual property (patents) and proprietary algorithms, we launched our sales activities in January. We have already secured one company for a PoC (Proof of Concept) and are discussing details with several others.

The implementation steps are very simple, with two options: API-type and package-type. With the API-type, companies only need to write a few lines of code into their existing website, and the AI automatically analyzes all pages to build the relationship graph and display the related cards.

Our pricing model during the PoC stage is set at "100 yen per page analyzed" plus a monthly system fee as we run validations.

For our growth strategy, we will start with "corporate recruitment pages," where results are most visible, then expand to general corporate websites and service introduction pages. Eventually, we plan to enter the fields of e-commerce (EC) and digital ad delivery. Integrating this intuitive discovery UI into EC and advertising will expand our market size exponentially.

While we are starting with B2B monetization to stabilize our revenue base early on, our ultimate goal is the B2C (consumer-facing) market. We envision services that leverage location data to let users intuitively discover art and culture info related to where they are currently standing.

I originally served as a business director at an art auction company. There, I saw many high-quality works of art that remained unsold simply because they lacked recognition. The algorithms of Google and modern generative AI are built to converge information toward the popular "center (majority)." However, this buries diverse, non-mainstream culture, destroying richness. We run this business with a strong passion to shed light on this "marginalized, unlit information" and provide opportunities to preserve a rich culture for the next generation. I look forward to discussions and meeting potential partners today. Thank you very much.

Q&A and Feedback

Mr. Nakazawa (Commentator): Thank you very much. While many AI tools focus on speed and time-efficiency (Taipa), providing an intellectual discovery experience where users can wander around productively is a very interesting approach. I strongly agree that searching (Search) and discovering (Discovery) should be treated as entirely different processes. My question is about the risk of mixing unnecessary noise for the user when presenting related information broadly. How do you construct this relationship graph or algorithm? What are the key details or techniques to display information with an optimal "distance"?

Mr. Mizuno: Thank you for the question. We pay close attention to the balance between weight scoring and user experience (UI). Specifically, the "top 4" cards displayed at the top of the Card-Style UI are filled with highly relevant, "convincing" information or items close to conversions. As the user scrolls down, we randomly place cards with unexpected related information that is slightly farther in distance—essentially "noise." If we present bizarre information from the start, users will leave. But placing an unexpected card right next to highly relevant ones triggers curiosity, encouraging spontaneous discovery.

Mr. Nakazawa: So you represent a gradation of information relevance through the UI design.

Mr. Mizuno: Exactly. We also incorporate an algorithm based on data "freshness." For example, yesterday's sports results lose value rapidly today. For such transient information, we have built a logic using the concept of a "decay rate" (cooling rate) to decrease relevance score over time. By combining these proprietary algorithms with our patented technologies, we create a system where users can comfortably wander and discover truly valuable information rather than facing a simple list.

Mr. Nakazawa: I see. Especially for future EC expansions, I see great potential in this system as it creates "space" (yo-haku) to increase purchase intent and trust in the brand. I look forward to your progress.

Mr. Mizuno: Thank you. We will continue to build a new internet experience that increases transparency and diversity, ensuring all valuable information receives the light it deserves. Thank you.