In the past few years, AI has greatly evolved and is starting to be adopted in various fields of business. Although AI is gaining a lot of attention, in what industries is it producing results? Sapporo Holdings Limited and Nomura Research Institute (NRI) worked together to start experiments in improving work efficiency at indirect departments using the AI system TRAINA. These experiments confirmed effects such as a 45% reduction in inquiry response work and an 80% reduction in information search time, and the project has now entered its full-scale deployment phase. We asked the people in charge at both companies about work reforms using new technology.
(Pictured left: Mr. Hidenori Kawamoto of the Group IT Management Division of Sapporo Group Management. Pictured right: Masashi Oishi of NRI Industry Systems Division 1. Department affiliations current at time of posting this article.)
An opportunity to tell upper management how you feel
Many companies face the problem that employees cannot allocate adequate time to their original jobs—such as visiting customers—because they are too busy with office work, such as handling inquiries and creating in-house documents. Hidenori Kawamoto of Sapporo Group Management says that the opportunity for this experiment came by thinking about how new technologies such as AI could be used to solve such problems.
"Sapporo Breweries Ltd. has a system that encourages employees to take challenges and gives them an opportunity to give a presentation in front of all the managers of Sapporo Breweries Ltd. after their thesis has been examined. I applied for the system because I thought it would be valuable to talk directly in front of the managers."
The presentation of evidence and definition of targets that take into account the status of both the market environment and the company, the system creation using AI, and the consistent proposal given by Kawamoto during testing, measurement, and subsequent operation were well received, and he reached the experimental stage. He was encouraged to continue by President Toshio Mizokami and Director Atsushi Ishihara of Sapporo Group Management.
Getting others involved in utilizing AI
In the experiments conducted between December 2016 and April 2017, a system for the centralized FAQ management for the in-house inquiry management site was created. Kawamoto focused on conveying the goal and output image to the people in charge so they could organize the FAQ data. AI enables the time needed to search for information to be reduced, increases search speed, and reduces inquiries themselves. Kawamoto thought that if the people in charge could realize the effects that AI can bring, they would promote on-going use of the system. FAQ management training was held in July 2017 in order to promote such perpetual FAQ updating.
"In order to effectively implement work reforms, people need to be motivated. I ended up rushing around the company to get everyone involved, and left system creation up to Oishi. The Group IT Management Division mainly works on planning and work proposals, and generally outsources system development. Since Oishi has been working on systems at Sapporo Holdings for many years and understands the situation, he was able to appropriately assign roles and the experiments proceeded quickly," says Kawamoto.
"Since AI differs from regular systems in that its precision increases as it is used, it is particularly important to involve other on-site employees at the company. Although it may not be completely successful at first, it is necessary to have a medium to long-term perspective," says Masashi Oishi of NRI.
Preparing to improve the productivity of the entire group
With the success of the experiments, Sapporo Group Management took the lead and decided on the full-scale adoption of the system. According to Kawamoto, "We currently have more than 600 items of FAQ data, which is more than was expected. The initial response accuracy was 60 to 70%, but after teaching employees the appropriate method for creating an FAQ and tuning the comprehension ability of TRAINA, the accuracy was increased to 80 to 90%."
Kawamoto has his sights set further than merely making indirect group work more efficient: he says that the system will truly gain meaning by deploying it in the group companies that handle beer and drinks, etc.
"In the 'SPEED150' long-term management vision, the Sapporo Group announced its goal of becoming a company with highly unique brands in the fields of alcoholic beverages, food, and soft drinks around the world by the year 2026. By utilizing this experiment, we will be able to make a wider range of proposals to our customers and this should increase the synergy between our different businesses. Since we originally wanted the system to aid our sales representatives, the real challenge is only just starting."
"Since the strength of TRAINA is its Japanese language analysis technology that can understand questions even if ambiguous words are used, display possible answer, and respond with questions to narrow down the results. The system can be deployed in a wide range of work types, such as sales, in addition to accounting and human resource management. The knowledge and schemes created together with this project can be applied to work improvement activities in any industry or company. If initiatives for increasing productivity can become more widespread, then I believe this can becoming something greater, something that can change the world," says Oishi.
It is difficult to propose work reforms and implement them while getting top management and colleagues involved. This project may be a useful reference when dealing with similar issues, such as its system for enabling individuals to express their thoughts, culture for assisting ambitious human resources, effective utilization of external resources, and appropriate allocation of roles, etc.
"Improving productivity and reforming work styles is a serious problem all over Japan. It is important to accumulate knowledge while cooperating with other companies, and I hope that our project can provide a good example for other companies," says Kawamoto.
Promoting on-site adoption of the TRAINA AI system
Kawamoto insists that for the on-site adoption of TRAINA to succeed, it needs to be able to do more than simply respond to inquiries. "I hope it can become able to give greetings and offer small-talk. It is important that each user can become familiar with TRAINA, as if it was their virtual secretary."
In response to this, Oishi says "That is actually the functionality we are currently implementing. We are training the system to respond to a statement such as 'You are cute.' with 'You flatter me.' or 'I'm tired.' with 'Let's have a drink together.' to give the system's responses a human flair."
Implementing measures to promote the utilization of TRAINA will accelerate work style reforms at the Sapporo Group.
Nomura Research Institute, Ltd.
Corporate Communications Department
Jul. 17, 2018