ExaWizards is a newly listed company growing into an integrated developer of AI products including their CareWiz line for solving social issues in the area of nursing and medical care. ExaWizards also provides AI integration services for industrial and social innovation, as well as exaBase AI platform offering one-stop-services for AI understanding, design, development, and usage, which has been implemented by over 500 companies including more than half of Japan’s top 100 companies and other global corporations.
We had the pleasure to talk with President Ko Ishiyama to learn more about the history of the company, the sophisticated AI solutions ExaWizards offer to support companies reach Society 5.0 and the wellbeing of their customers, as well as their unique multi-sector and multi-modal strategies.
Western media heavily criticizes the Japanese private sector, saying that it lacks innovation, citing the small amount of startups or unicorns that are present in Japan, and the corporate rigidity. Do you agree with that criticism of Japan? And as an entrepreneur yourself, what do you think is the innovative and competitive edge of Japanese industry today?
Firstly, the Japanese government started to try to introduce the concept of “digital garden city state initiative” using DX or IT, and one of the issues surrounding it is - when Japan introduces IT or digitalization, how can that fit into a society which is centered around people?
“Society 5.0” is one of the words or phrases used in Japanese policy around the science and technology it has been utilizing so far. Since Exawizards is a company trying to solve social issues through AI technologies, we received funding from INCJ, an investment fund majority owned by the Japanese government, and then did the IPO process.
In my understanding, our AI or computer science is in the third paradigm shift – the world’s AI or computer science is in the third paradigm. The first paradigm is a kind of competition paradigm with regards to the research around science and technology and IT. The Chinese Singhua University tried to increase the number of people who are working in that field.
The second paradigm was an era, or a period when GAFA in the US and Alibaba in China tried to utilize AI in the business field. The third paradigm is about society. For example, in the case of the US, there is an issue around their divided society because of fake news. Facebook ended up in a situation where they needed to rename their company because of this.
In China there is also an issue around their surveillance society. What does this mean in reality? Even though people tried to publish many academic papers to try to utilize their expertise in the business world, there has been no company in the world at the moment where people have successfully been able to reflect their social values.
What kind of value can we generate in society through this third paradigm? That is very important in Japan, I think, and I would like to share an interesting example. About four years ago we received a request to explain about the cashless system in Japan to a student who came from the Chinese People’s University in the MBA course. At that time, we were wondering what we could talk about with people from China because they had been far more advanced in terms of cashless compared to Japan, so we didn't have anything to say. We devised a kind of story that we could quickly tell them about a presentation a Japanese delegation made to be able to host the Olympics in a big city in the past. A popular concept they utilized at the time was omotenashi in Japanese, so we tried to explain what that is.
At that time, one of the examples they utilized to explain about the hospitality or omotenashi in Japan was a story that even though a person might drop a wallet with millions of yen in it, they can finally get all the money back in Japan. By saying that, we asked whether or not China can also get back the amount of money with an advanced cashier system.
What we were saying is that Japan has been far advanced compared to other countries because even though we don't have advancement in cashless systems or policy, we can get the money back anyway. That is what the Society 5.0 is all about. That's what we said. We are working on the use of AI to realize such a people-centered society.
Japanese companies are changing. Under the Covid19 situation, the demand for a variety of goods is increasing and logistics is becoming more complex. Companies need to provide sustainable service, and demand forecasting has become a key factor.
Yamato Transport, a major logistics company, is using a method called MLOps to forecast demand and optimize costs. MLOps is a continuous improvement method of machine learning algorithms used by big tech companies like GAFA, and ExaWizards helped Yamato Transport implement it.
Thus, ExaWizards is not simply introducing MLOps to companies, but is automating the solution of social issues through machine learning.
The Japanese population has the oldest average life expectancy in the world of 85 years. More than 1/3 of the population is over 65, which poses a reduced labor force and less demand for products in general. How do you think that Japan’s aging society will impact the development of AI technologies, and what is your firm doing to use this technology and assist Japan as it emerges into this aging system?
I believe that the super aging society is one of the biggest challenges or problems we need to solve through AI in society. These are actually the words uttered by Marvin Minsky, the father of AI. Until he passed away, he was actually in a position to be an advisor to ExaWizards.
We have been trying to further enhance our AI technology thanks to him. He used to say that we should solve social issues through AI in a super aging society and I believe that Japan actually has advantages because we will face this problem before the others.
In 2019, Nobel Prize Dialogue Tokyo was held. That is a venue where experts formed a panel to discuss this topic with Nobel laureates including various professors and myself as well. Almost all the Nobel laureates said at that time in the venue that Japan is going to be the fastest aging society in the world so the research done in Japan should be shared with other countries.
With that, I would like to start talking about what kind of AI we have been doing to be able to deal with the problem. There are two main issues in an aging society. Social insurance costs will be increased because of the increasing number of elderly people and also the decrease in the working population.
To be able to deal with these problems we developed exaBase, a DX platform to be able to somehow deal with the issues around the decreasing labor population and we also have an issue with nursing care for increasing social security.
In 2040, Japan is expected to experience a decreasing number of labor forces by about a few tens of thousands of people. In Japan we have strength in our manufacturing industry, however, we already recognize that we're going to have a significant shortfall in the labor force that has been underpinning the plant work in that field. For example, we need to think about the automation of factories. Let's say we have parts and components for cars and they have to be inspected visually by people manually. By utilizing exaBase we can use robotics to be able to do that instead of manual operations.
One of your partnerships is with Aflac, a very large insurance group. What have you learned from this collaboration with Aflac that you could apply to the entire insurance industry or that can be taken to other fields?
We’ve actually already reached partnership agreements with another two insurance companies in addition to Aflac. Under the current circumstances of a super aging society that we are in right now in Japan, of course it's important that the government has been providing public insurance to people but also we believe that it's important that private insurance companies have to be grown too.
In particular, I can say that the commonality among any insurance companies that we have been working with is that they try to increase or enhance customers’ well-being. If they can increase or enhance customers’ well-being, they can increase their sales productivity. In the past, they used to say that when the NPS (Net Promotor Score) of customers increased, their employees’ score would increase accordingly so it's not simply about the satisfaction level among customers only.
When we think about expansion from NPS we start to see a correlation between customers’ well-being and employees’ well-being. We’re entering an era where insurance companies are starting to sell more services involving the enhancement of customers’ well-being together with their insurance products.
As for what’s happening with DX, AI is starting to contribute to the enhancement of well-being. For example, what kind of sales or sales-related activities should be changed so that they can successfully enhance customers’ well-being, which will end up with the enhancement of our own well-being. That’s what AI is starting to analyze.
Going back to robotics, it's important that we’ve started to see more advancement or evolution of automation in factories. It's particularly important that we need to transfer the skill set or a caliber or capacity of skilled workers in manufacturing craftsmanship that Japan is so proud of to the next generation.
ExaWizards developed a robot which can reproduce the capabilities of skilled veteran workers in manufacturing craftsmanship. The other day we just published some PR information in order for us to showcase what we can achieve in the robotics sector because we are planning to show a brief demo at an international robot trade show.
We developed AI which can successfully present a pancake decoration, and not just a simple representation of a pancake in general, but our robot can achieve the kind of pancake decoration done by a Michelin chef. At first we asked veteran chefs to have a look at several of images of pancakes to make a judgment about whether each one was good or bad, and we would analyze that data.
After that, the AI started to automatically generate data to say which decoration was better done by chefs and what kind of decoration they should have on what kind of pancake. The AI could then not only just create or reproduce the image of a beautiful pancake, depending on the type, the AI can also decorate exactly what it's supposed to be. Actually, it's very difficult to put a blueberry on top of a pancake as a finishing touch but we developed a robot which can successfully do that. Other examples include adding raisins on top of something but actually the factory has so many tasks that have to be successfully done one by hand and our AI robots can do that.
Looking at the future, are you looking for new partnerships and if so, what new applications do you foresee for your company?
We have been adopting very rare strategies called multi-sector and multi-modal. Examples of the multi-sector approach include Alibaba or Shopify, who started with e-commerce and then moved on to payment businesses and then moved to renting and then to the insurance business, so data is very disruptive and then can be utilized beyond the boundary of sectors.
We also have interest in the nursing care sector, we can provide AI services to insurance involving dementia to insurance players, among others. As another example, when we watch the working pattern of elderly people, we can gather data and suggest that a person might be suffering from a genetic related disease or symptoms so that a specific type of medicine needs to be taken.
Additionally, beverage companies sometimes want to sell non-alcoholic beer to nursing care facilities. There is evidence that dementia symptoms might be weakened when there is improvement in the intake of water. This is exactly the essence AI can generate because of a kind of crossroad of the good points or advantages of those sectors within the multi-sector.
Next, I'd like to talk about the multi-modal. There is not only one pure standalone data set, but multiple data sets have been integrated together. For example, when we have a meeting with the customer, video data will remain because of that. When we communicate through chat, text data will remain as well.
Structured data can be obtained about, for example, what kind of sales activities or approaches we have taken to specific customers today, and when they have a verbal conversation, we also need to make an analysis about the voice data.
In order for us to enhance the well-being of end customers, we analyze all the things that I said earlier, and then finally we can recognize what kind of things can be contributors to finally increasing or enhancing well-being.
When it comes to social issues or challenges, they are basically abstract things. It means that multiple data sets have been combined or intertwined together so we need to handle everything otherwise we cannot achieve the best results.
We are actually not running a business driven by technologies, but we have been running a business driven by problem solving and as I mentioned, because we are in the third paradigm, we haven't just tried to utilize what we have developed as a technology, but we try to solve challenges that society has been facing and look for the kind of social values AI can generate to be able to run our business.
Can you tell us a bit more about your international strategy and how you plan to expand the company beyond the Japanese border, both in terms of financing and in terms of operations?
In Japan we have a shortage of digital talent. In order for us to launch AI related startup companies, it's actually impossible for us to be able to do that with only Japanese people. In our company, about 30% of our engineers are foreign people.
For example, we can think about the combination of a foreign engineer and Japanese people who are specialized in the nursing care business or services. Through that kind of new combination we believe that we can deliver new services.
That is the kind of thing we can utilize to solve social problems and that is a strength of our company. We are facing a super aging society ahead of other countries and that is an advantage we could utilize on top of that combination of domain specialists in specific areas and foreign engineers.
We chose ExaWizards as the company name because for example, nurses who are experts in the nursing care business or services for the super aging society in Japan are also some of the wizards, and machine learning (ML) experts globally are some of our workers as well. That's why we put Wizard in the company name.
That was the first step and the second step in our super aging society in Japan which we can achieve in Japan ahead of other countries can be reflected in or expanded to other countries. That's what we are trying to do in the future.
Let's say we come back to interview you again on the last day of your presidency. What would you like to tell us about your goals and dreams for the company by that time, and what would you like to have achieved by then?
Let's say that the Japanese government or state has been dealing with about 20 trillion yen in terms of social insurance costs for nursing care insurance. Through AI, we believe that 20%, or four trillion yen, can be saved. Of that amount, we think that we can make an improvement of about 20% as Exawizards, so that would be about one trillion yen. That's what we'd like to earn as part of our revenue.