Hmcomm ＜265A＞ visualizes sound by utilizing artificial intelligence (AI) and provides services to solve customer issues. With the acquisition of an AIST technology transfer venture, AI research and development focusing on “sound” has been carried out, and the results have been aimed at implementing the results in society. It is a B2B business where companies with voice recognition and noise detection needs are customers, and although there are cases where projects go through agencies, 80% of company-wide sales are direct contracts. Research and development of core technology and development of related upstream processes are handled internally, and an efficient and quick development process is carried out while maintaining technical capabilities by dividing roles in the form that outsourcers are responsible for downstream development processes other than core technology. The segment is divided into an AI product business (69.5% compared to sales for the fiscal year ending 2023/12), which develops products that can be used for a wide range of customer groups, and an AI solution business (same 30.5%) that handles custom-made services tailored to specific customer needs.

AI products include an AI voice recognition product “Voice Contact” for call centers, an AI automatic voice response product “Terry,” an AI meeting minutes product “ZMEETING,” and an abnormal sound detection product “FAST-D.” Compared to noise detection, the ratio of voice recognition is high, making it a recurring revenue model. In the third quarter of the fiscal year ending 2024/12, the number of AI product suppliers (number of companies) was 41 companies, and the average customer transaction unit price was 12.3 million yen. Meanwhile, the AI solution business is a contract type non-recurring revenue model, the number of projects is 46, and the average unit price per project is 5.2 million yen.

Sales for the third quarter of the fiscal year ending 2024/12 landed at 744 million yen and operating profit at 125 million yen. For AI products, the average unit price has risen for 2 consecutive terms (1.7 million yen increase compared to the same period last year). The number of accounts was at the same level as the second quarter, and it seems that spread within the same account was seen. Also, in the AI solution business, in addition to continuing projects from the previous fiscal year centered on DX, it seems that new research and development projects related to large-scale domestically produced language models were also received and progressed smoothly. Full-year sales are expected to be 1040 million yen, up 29.9% from the previous fiscal year, and operating profit is expected to increase 55.9% to 129 million yen.

The company's strength is that in addition to carrying out AI research and development focusing on “sound,” it can complete everything from research and development to product development in-house, and it is promoting the transfer of learning models constructed starting with problem-solving research and development to the next problem solving. Since it is possible to handle everything from AI solutions to AI products with a unique research and development type business process, it is compatible with a wide range of companies, mainly small and medium-sized enterprises, and can be flexibly customized according to customers, and differentiation from other companies is achieved by providing services close to customers. For example, Yaskawa Electric (6506) automates shipment determination (sound detection) work at the factory, and with Benesse Holdings, AI “unmanned” the “reception/completion, and order acceptance operations” of mail order acceptance through natural conversations with customers. Furthermore, JR East ＜9020＞ has also been able to build co-creation relationships with large companies, such as systematizing defect judgments in railway business assets from the viewpoint of noise.

According to a third party agency, the domestic AI market is predicted to grow at an annual rate of just under 9% over the next 5 years, and it is assumed that the market size will remain steady even in the voice recognition/noise detection area. Above all, there are observations that predictive detection solutions will approximately double in the next 5 years. Going forward, the current number of AI product introductions per company for AI products is 1.1 (fiscal year ending 2023/12), and the number of AI products per company will be raised by strengthening cross-selling and expanding the number of AI products. Also, with the aim of accelerating user acquisition, it seems that they will also seek to build up the number of agents in the future. While expanding AI solutions, which serve as leading indicators for AI product development, stock sales of AI products will also steadily build up. It seems that they will promote research and development type business processes, expand the top line through the growth of AI products, and aim to become the No. 1 company in the “sound x AI” specialist group. As steady trends in business performance continue, I would like to keep an eye on future trends as a company specializing in “sound” even among AI-related recent IPO brands.