Vol. 78, No. 4, July 2023

KOYAMA Hiroyuki

HAMADA Mikio / TAKANASHI Mari

Toshiba Tec Corporation is aiming to transform itself from a business that sells products to a platform business, starting with its point-of-sale (POS) business in which it holds a large market share. We are working to expand our customer base not only in the retail industry but also in related industries such as payments, advertising, and logistics, and to construct an ecosystem that provides true value, thereby contributing to the realization of the Sustainable Development Goals (SDGs) and enhancing people’s level of satisfaction in their lives.

In order to transform into a platform business and achieve further expansion, we established an organization of service design specialists in October 2022 to lead co-creation activities through dialogue and empathy with stakeholders. In addition, the organization is developing original methods for service design and business studies, and providing opportunities for co-creation by means of dialogue with customers and consumers, problem discovery, and idea generation. Through the development of human capital, opportunities, and methods, we will continue our efforts to create a system that comprehensively supports the provision of value to society.

ISAKA Takehiko

The realization of a sustainable society has become a significant international trend accompanying the economic development of countries around the world. In this context, it is necessary for the retail industry to shift its concept of corporate activities from ownership to sharing, recycling, and reclamation, and to increase the efficiency of resource utilization.

The Toshiba Tec Group is committed to enhancing satisfaction in people’s lives by providing retail businesses with components of its services so that they can solve economic, social, and environmental issues from multiple perspectives, including in the areas of socioeconomic systems, lifestyles, and technology.

KURODA Kazuyo

The current market is characterized by the release of a wide variety of services in rapid succession. In this environment, it is becoming difficult to achieve differentiation in the conventional sell-out process, in which the specifications of a service are examined based on the manufacturer’s hypothesis and the service is then developed and released.

Toshiba Tec Corporation has studied this issue and developed a service design process that can swiftly and continuously provide value to customers. This process involves modeling the systems required to provide a service and repeating the Plan-Do-Check-Act (PDCA) cycle for that service from the customer’s viewpoint based on a service design method. We aim to apply this service design process to the creation of differentiated services that provide true value to customers.

TERANOSHITA Masahide / KAMISUWA Yoshikatsu / HARA Noriyoshi

In a situation where customers’ business environments are changing at an accelerating pace, it is necessary to swiftly respond to the issues encountered by customers through dialogue and co-creation with them. For this purpose, it is important to design the required services based on post-release information collection and feedback to the customer.

Toshiba Tec Corporation is constructing a service operation platform that supports the management, analysis, and status assessment of equipment by centrally managing the operating status of all systems and equipment installed by the customer in real time and visualizing the status in the form of a bird’s-eye view. The continuous management of services in this way will make it possible to reuse and recycle services and contribute to realization of the Sustainable Development Goals (SDGs) through the reduction of development resources.

TAKAHASHI Nobuyuki / HARA Noriyoshi

The retail industry is facing a chronic labor shortage, and while workstyle reforms and automation are being implemented, further issues have emerged resulting from the impact of the coronavirus pandemic, soaring prices due to international conflicts, and the depreciation of the yen.

Toshiba Tec Corporation follows the principle that not only actions taken by the company itself but also collaboration with its business partners is indispensable in order to promptly provide services to solve the various problems being encountered by customers. For this purpose, we have constructed an ecosystem for new value creation through co-creation and collaboration with our partners, and are providing a support system that enables us together with our partners to quickly develop solutions utilizing the ELERA global retail platform.

KOBAYASHI Daisuke

Technologies for detecting specific objects in images using deep neural networks (DNNs) and other techniques are becoming increasingly important to improve productivity, promote automation, and save labor. However, retraining of a DNN is necessary in order to detect new objects that did not exist at the time of its introduction. The few-shot object detection technique makes retraining possible using only a few images, but this process takes at least several hours even when high-speed hardware is used.

Toshiba Corporation has developed a few-shot object detection technology using feature vector registration that requires only a few images of a new object together with the correct information, eliminating the need for retraining. We have also developed and applied an original self-supervised learning method that automatically learns the shapes of objects in an image other than those targeted for training, improving the accuracy of detection of new objects compared with conventional methods.

KOBATAKE Tomoya / YAMADA Masataka

In order to maintain the performance of artificial intelligence (AI) applied to the provision of actual services, where data are subject to change due to various factors, it is necessary to continuously update the AI model. However, the updating of an AI model consists of highly specialized processes such as monitoring, relearning, and AI model management, requiring substantial human resources.

The Toshiba Group has developed a machine learning operations (MLOps) platform that automates these processes and allows them to be smoothly implemented on site, thereby improving the speed and quality of development of AI-based Industrial Internet of Things (IIoT) services. We have applied this MLOps platform to IIoT services such as plant anomaly prediction and automatic product inspection, and confirmed that it is possible for personnel who are not AI specialists to execute the process of updating AI models and to take the necessary measures in response to changes in data.