TM-RoBo Trademark Search Function Enhanced to More Deeply Reflect Application and Registration Status

IP-RoBo Co., Ltd. (Head office: Minato-ku, Tokyo; President and CEO: Masafumi Iwahara; hereinafter "IP-RoBo") provides "TM-RoBo," an artificial intelligence service that supports trademark searches, including the registrability of trademarks and the risk of infringement through use. We have added a new feature to TM-RoBo's trademark search function that more deeply reflects the application and registration status related to the trademark being searched, and this service will be available starting April 1, 2026 (Wednesday). This new feature will significantly improve the accuracy of word segmentation, distinctiveness indicators, and the final comprehensive indicator, making it possible to conduct trademark searches more easily, quickly, and accurately than ever before. ■ Background Our company released "TM-RoBo" in April 2018, an artificial intelligence that quantifies various indicators useful for judging the registrability of trademarks, by machine learning the judgments of experts such as the Japan Patent Office and courts. Since then, we have aimed to dramatically improve the efficiency of trademark search operations by sequentially adding new features. In contrast to conventional systems, which could only search by inputting a series of katakana pronunciations as a single unit, the released in July 2020 is a groundbreaking feature that realizes the following functions: 1. By inputting a word mark as it appears (using any character type such as Kanji/Hiragana/Alphabet as is), designated goods/services, and similarity group codes, the AI assigns pronunciations and calculates word power statistical indicators (*1). 2. When a compound trademark consisting of multiple words is input, the AI comprehensively extracts all separable combinations of words, then analyzes each combination in detail, calculates various indicators such as TMR (*2) and TMC (*3), and finally calculates TMS (*4), which is the comprehensive indicator. 3. If necessary, it searches for and displays examples of coexisting registrations useful for judging the recognition of essential parts. *1 An indicator of the relevance between the target word and the target goods/services, related to distinctiveness. *2 An indicator related to how original the pronunciation of the target word is in relation to registered trademarks, etc. *3 An indicator related to the possibility that the target combination of words may be observed separately. *4 A comprehensive indicator regarding the safety of the target combination of words or the trademark under investigation. ■ Challenges Regarding the accuracy of the word power statistical indicators related to distinctiveness, we have improved accuracy by adding numerous appeal decision data and our own proprietary data. However, due to the vast amount of data required, there were cases where inaccurate results were output. In response to the above, we have improved accuracy by sequentially adding developments as follows: 1. February 10, 2022: Added word power statistical indicator data by focusing on specific fields. 2. October 1, 2024: Added examination stage data and examination standard data. In particular, the addition of examination stage data mentioned in point 2 significantly contributed to improving accuracy, as the amount of data increased several times compared to before. However, there were still a considerable number of words that did not appear at all during the examination stage, and improving the distinctiveness of these words remained a challenge. ■ Function Overview To solve the above challenges, in addition to the conventional approach of collecting data judged to lack distinctiveness, we decided to add an approach that deeply analyzes trademark registration status. Specifically, we have implemented the following functions: * (1) More detailed searching and deeper analysis of registration status related to the trademark under investigation. * (2) More accurate calculation of word power statistical indicators (*1), taking into account the analysis results from (1). * (3) More accurate proposal of word segmentation positions, taking into account the analysis results from (1). * (4) Provision of materials to respond to office actions (refusal notices), etc., by displaying the analysis results from (1). * (5) Calculation of a comprehensive indicator for usability surveys, in addition to the comprehensive indicator for registrability surveys. ■ Effects This new feature will significantly improve the accuracy of word power statistical indicators, which are related to distinctiveness. Furthermore, it will strongly support not only surveys regarding the possibility of acquiring trademark rights but also usability surveys to determine whether a trademark can be used safely without infringing on third-party trademark rights. More detailed data regarding AI indicator calculations will be provided than ever before, making it easier to check the reliability of each indicator and also contributing to the enhancement of evidence gathering when responding to office actions (refusal notices), etc. ■ What is TM-RoBo When newly using a trademark as a name for goods or services, or when attempting to acquire trademark rights, it is extremely important to investigate whether identical or similar trademarks have already been applied for or registered. Such trademark searches are commonly conducted. However, existing systems for trademark searches make it difficult to perform more crucial similarity searches, as they can only conduct exact match searches or similar searches when searching for the trademark as it is written (trademark search). Furthermore, while similarity searches are possible when searching by pronunciation (how a trademark is read), the results cannot be displayed with the degree of similarity taken into account. This meant that investigators had to painstakingly check all the large number of listed trademarks. For these reasons, trademark searches required not only high-level expertise but also considerable effort and cost. Our company released TM-RoBo in April 2018, an artificial intelligence that machine-learns the judgments of experts such as the Japan Patent Office and courts. Since then, by sequentially releasing additional features, we have supported the dramatic improvement of trademark search operational efficiency using the latest AI technology. In addition to the newly released registration status reflection function and combined word expansion function, the current TM-RoBo also offers the following features (some are paid options): * AI Pronunciation Search Function * Similarity Group Code Search Function * Similarity Group Code Template Function * Batch Pronunciation Search Function * Display Item Addition Function * AI Trademark Search Function * Word Power Statistical Indicator (*1) Calculation Function * Appearance Partial Match Search Function * Pronunciation Partial Match Search Function * B+A Type Search Function (Conceptual Similarity Search Function) * Self-Registration Function * Batch PDF Output Function * System Integration Function etc. Official Website: http://www.tm-robo.com/ ■ About IP-RoBo Company Name: IP-RoBo Co., Ltd. (http://www.ip-robo.co.jp/) Representative: Masafumi Iwahara, President and CEO Location: Shinbashi TS Building 7F, 1-22-5 Nishi-Shinbashi, Minato-ku, Tokyo Established: March 20, 2018 Capital: 10.67 million JPY Business Activities: 1. Consulting business (general consulting related to intellectual property) 2. Web service business (internet information services related to intellectual property) 3. Software development business (design and development of software related to intellectual property) Contact: Mail: info@ip-robo.co.jp