Only a decade ago, people thought of AI as something from the far-distant future, yet today it is a firm fixture of everyday life. Our smartphones contain technologies underpinned by AI: voice recognition, auto-calibration in our camera apps, and translation tools. AI is now taking hold in the world of business, with products emerging in all sorts of industries from manufacturing and finance to communications and logistics.
The retail stores we use every day are another part of our lives where automation is becoming more popular. Alongside its applications in customer service and product management, stores are increasingly using security cameras for purchase analysis.
AWL is a company that specializes in AI camera solutions. The company is developing products that offer automated image analysis from security cameras installed in stores to facilitate store operations.
AWL President Muneharu Kitade has said that, “My dream for the future is realspace data conversion that helps companies to optimize.” How did Kitade end up in the AI business? We spoke about the leadup to beginning the company and his vision for the future.
・Good technology can change people’s lives. I felt that potential in AI.
・AWL was born from the worry that if we tried to offer everything, we would win nothing.
・From retail to the office: Using the accelerator program as a springboard to enter new sectors
・The secret to successfully working with big companies is enthusiastic collaborators
・My goal is to make society more convenient through realspace data conversion
Born 1978 in Tomakomai, Hokkaido. Kitade launched an online business while at university and following graduation, went on to work in web marketing and consulting at a US consulting company (DC) and a US record company (NY). After coming back to Japan, Kitade helped to establish the IT department of GABA, a company providing one-to-one English instruction. After building experience in numerous business launches as media operations manager of Livedoor, he left to become independent in 2006. Kitade works on business and company startups and provides consulting to major companies. A meeting with Professor Kawamura of Hokkaido University pushed Kitade toward societal applications of AI, leading him to found AWL, Inc. (formerly AI TOKYO LAB) in June 2016.
Good technology can change people’s lives. I felt that potential in AI.
Kitade’s background was mainly in business startups. Before launching AWL (formerly AI TOKYO LAB) he had not come into contact with AI. Why did he decide to get involved with this sector?
Kitade: The catalyst was a business request I had from a client in 2015. This client had asked me several times about their interest in using AI to solve business-related problems, and I was looking for a collaborator with whom to work on a proposal. I got in touch with one of my company’s clients, who told me they knew just the right person and put me in contact with an AI researcher.
That researcher was Professor Hidenori Kawamura of Hokkaido University, who is now a technical advisor to AWL. Kawamura explained what was happening in AI technology and its technical potential to Kitade, who thought that AI would be hugely important in the future.
Kitade: I think good technologies can really change people’s lives. Listening to Professor Kawamura, I was convinced that AI would absolutely become a key technology that changes the way people live.
Actually, my own life and my values were changed when I discovered the Internet during my time at university. My grandfather ran a wholesale market, and ever since I was little, I had thought I had no other option when I grew up other than to take over the family business. It’s a tough life in the wholesale markets; you have to get up at four in the morning even in the middle of winter to prepare for the auctions, and there’s a lot of manual labor involved. I’m not physically strong or a morning person. The idea of one day taking on the running of the market really made me feel miserable.
My discovery of the Internet turned that hopelessness on its head. With the Internet, you can transcend time and space to make so many things possible. If AI becomes similarly widespread, then manpower-reliant jobs like cash registering and warehouse management could be taken over by technology. Even though you might be told to mind your own business, you can tell people who think their capabilities are limited to the work they are doing that they have the potential to do something else.
With this sense of AI’s potential, Kitade teamed up with Professor Kawamura and they started working toward commercialization. Starting with a client request to automate promotional material proofreading, the pair launched a business in 2016 that developed AI-based services on a contract basis.
Kitade: One client was producing multiple catalogs every year. Their copy editors were getting old and it was hard to pass on their knowhow to the younger generation, so our challenge was to find an AI solution to this problem. If you’re going to do business, you need to have a company that can take on the work. We rushed to launch AI TOKYO LAB, the predecessor to AWL.
At this point in time, Kitade’s approach was to give the venture a go since there were hardly any examples of AI success, but the company found itself in great demand. It was inundated with requests for different AI development projects including character recognition, voice recognition, image recognition, and numerical simulations.
Kitade: We were more like a project team than a company at the beginning. I had been involved with 10 or so startups since becoming self-employed 10 years ago so my approach was to give it a go and see what would happen. I felt that even if I failed doing something I didn’t understand, I would still gain knowhow and develop my skills. But the world of AI was much deeper than I imagined, and there were moments when things were really hard.
We had loads of researchers on the development side, so it was hard to get everyone on the same page. With all the schedule delays and abstruse work full of jargon they were producing, I couldn’t believe how different values were to the business side. There were some conflicts between the two sides, but with the support of Professor Kawamura, I was able to keep the business and development sides in line and working together.
AWL was born from the worry that if we tried to offer everything, we would win nothing
The company’s business was steadily growing, but within a few years of its founding, other players had begun to enter the AI market. Kitade began to grow strongly concerned that the business would have no future if they remained complacent.
Kitade: When we first launched the business, we always had enough work just from all the AI development we had the capacity to do. However, other players were getting involved in the sector and there were more and more companies appearing with a specialist niche. I remember thinking that if we carried on being AI generalists, we would end up fading away.
AI technologies are rapidly advancing all the time. If you don’t hone in on one area, it will be difficult for you to catch up, and if your developmental capabilities fall behind, you will end up less competitive. What’s more, some of our engineers were telling us that they wanted to work on solutions that would be unique to the company.
However, Kitade wasn’t sure which area to focus on. Then came a request from a client that provided the impetus to move into AI cameras.
Kitade: We had some conversations with Satudora Holdings, the company operating the “Satudora” drug store chain, about how the use of AI cameras could improve their stores. As we listened to them describe the issues and got to work on solutions, I gained the sense that this was an area that had huge potential.
At that time, there were only a few players working in realspace data conversion, and none were going beyond the trial experiment phase. But if we could find a way to link up with IoT and put the data to good use, I knew we could deliver efficient store management. Moreover, unlike in digital spaces like apps or web products, it would be easier to create an image of the user because the system would be used in real stores that people visit. The engineers were also excited about the challenge.
Thinking this was an area with potential, Kitade began to put feelers out among his clients in the retail industry. While he didn’t yet have the product, he found that when he described the services he was envisaging and how much it would cost, many of his clients said they would definitely purchase the product at the cost level he outlined.
Kitade: From these conversations with clients, I gained a grasp of the common issues in the industry and the rough price that companies would be willing to pay to get hold of the service. I felt that sticking with a generalist offering could be bad for the business, so it was worth taking a risk and specialize in realspace data conversion, which led me to focus exclusively on the development of a solution called AWL (AI + OWL: Owl origin), with the company name changing to AWL as well.
I decided to sell the consulting business I had developed under the AI TOKYO LAB brand in 2019 in order to focus our resources on the new venture and launched the new brand AWL to set off in a new direction.
From retail to the office: Using the accelerator program as a springboard to new sectors
After the company name change, AWL refined its focus to the retail sector and got to work developing an SaaS product. The idea was that by connecting AWL devices to security cameras and signage in stores, users would have access to various functions including purchase analysis and antitheft/out of stock alerts.
Kitade: Basically, all retail stores have security cameras so that they will have evidence if there are any incidents. We developed AWL’s product in a way that utilized security cameras – assets that stores would already have. Simply by connecting AWL’s system to their security cameras, users would be able to access AI functions like purchase analysis and antitheft alerts. The use of an existing asset meant we could keep down installation costs and could maintain operations at a tenth of the monthly cost of our competition.
Budgets are tight in the retail industry, and companies are unwilling to make new budgets. But since we were able to say that our product would actually cut costs and would allow our customers to make better use of their existing security cameras, we steadily managed to increase the number of users.
AWL then joined Mitsubishi Estate’s accelerator program in January 2020. The program helped AWL accelerate its offering. Where it once only provided solutions for retail application, they are now developing solutions for the office.
Kitade: Image analysis technologies can be used wherever people gather, whether that’s in the office, factory, hospital, or school. However, it’s difficult to scale a product in a specific sector until you’ve worked out how it fits into that sector. During the company’s early days, our focus was on retail, but I was keen to enter other sectors as well.
Just at the point where I was thinking about the office as the next sector for us to enter, I heard that the major property manager Mitsubishi Estate was running a program. I applied and was lucky enough to be accepted, so I moved our head office to Shin Tokyo 4TH, a new office managed by Mitsubishi Estate in the Marunocuhi Shin-Tokyo Building which was set up to encourage company DX and open innovation, and that’s where we started to develop office-oriented products.
During the program, AWL developed systems for the body temperature detection and facial recognition of visitors, and lounge and smoking room congestion detection. People working in a building can check its congestion levels on their smartphones, helping to avoid the Three Cs (closed spaces, crowded places, and close-contact settings).
Kitade: The Coronavirus crisis and the accelerator program was a turning point in which we took a hard look at the business.
Most of the existing business negotiations for our retail store-focused system came to a stop because of the crisis. But there were also all sorts of new customer needs, to avoid queues, busy places, physical contact, and so on, so we quickly started developing Coronavirus-related functions. These created new business opportunities for the company; although our sales composition ratio changed, we were able to maintain the same level of performance as before.
Being part of the program during development had a major impact. AWL knows how to develop its own systems, but we don’t have anywhere to implement and test them. Thanks to Mitsubishi Estate providing us with the office space where we could carry out trial experiments, we could move frictionlessly ahead with development.
The Coronavirus crisis really affected the company, but we avoided disaster by responding nimbly. With the spread of remote work that we’re seeing now, interest in digital transformation is only going to grow and I think that will really help the business.
The secret to successfully working with big companies is enthusiastic collaborators
The fact that AWL built a new product through the accelerator program is a great example of the program’s success. However, collaboration between startups and big companies can often be bumpy because of the differences in company culture. What did AWL do to ensure smooth collaboration?
Kitade: AWL has been selected for accelerator programs three times in the past, and it’s chance that determines whether the collaboration will go well. There are certain things outside our control, so there is an element of having to try it and see what happens.
Obviously, any startup needs to produce clear materials and videos to facilitate meetings. But there’s no killer move; sometimes you get good synergies, sometimes you don’t.
So why did the collaboration go so well this time? Kitade puts it down to the proximity of physical distance and the enthusiasm of those involved.
Kitade: This could just be hindsight talking but moving our headquarters to an office owned by Mitsubishi Estate really helped. During the implementation stage, we were in an environment where we could ask to have an immediate meeting or say we were bringing over the device right then and there, and this allowed us to keep up an enthusiastic momentum.
The enthusiasm of our contact at Mitsubishi Estate was fantastic. The guy we worked with made sure our requests were passed in-house. He had transferred to his current department due to his interest in startups. He said it was his job to help us and it was him being proactive about it that meant we could bring in the cameras and systems so smoothly.
Kitade says that having the right person in the role during collaboration is really important. When a big company is onboarding a system, internal coordination is necessary, and the process needs to be signed off by management, IT, and the legal department. Their contact at Mitsubishi Estates took the lead in acting as an intermediary and introducing Kitade to decision makers. The enthusiasm of the individual is what drove the big company forward.
My goal is to make society more convenient using realspace data conversion
Finally, we asked Kitade about his vision for the company. What sort of world does he envisage building through realspace data conversion?
Kitade: It’s hard for me put it into exact words, but what I want to create is a world where AI augments human creativity and potential.
The advantage of digital transformation is that it’s easy for business owners to understand. If you tell a business owner that introducing unmanned cash registers will allow them to cut their register staff of 600 people, they’re going to be delighted and say they want to do it. But then one of those people will have worked on the cash registers part-time for the last 20 years to make their living, so they will see you as stealing their job. I can see where those people are coming from.
But you know, technology stops for no one. Hundreds of years ago we used to get around by horse or ox, but then we created cars and airplanes; jobs come, and jobs go. The moment in time where a type of job is lost is painful. But new types of jobs are constantly being created. Perhaps the new jobs will offer opportunities for people to realize their new value. There are labor shortages in all industries, so they are going to have to update their valuable human resources. A safety net will be needed, but the growth of AI is unstoppable.
How will people work together in a future where AI is used everywhere? Kitade talked about how Amazon Go is a perfect example.
Kitade: People’s perception of Amazon Go is of unmanned stores, but in reality, there are people stationed there. There are the staff at the entrance helping you to enter the store who chat to you about the weather and provide a warm welcome. If you go to the same store all the time, they will remember your face, and there’s such a high level of hospitality, it’s like going to a luxury hotel.
There’s also the staff at the alcohol section who check customers’ ages, and they’re always friendly. They act like sommeliers, asking you what’s on the menu that day and offering suggestions for certain wines that will go with your food. Of course, there are going to be people working in the store, and they’ll make you want to go back there.
My vision is to have a hybrid of people and technology that is achieved through realspace data conversion and labor saving. Rather than have completely unmanned stores, I want to create a world where AI is used as an assistant to help people offer extra value.
Kitade says his vision for the future is to use realspace data conversion to make society more convenient.
Kitade: The social implementation of AI cameras is controversial. Some people think they might lead to surveillance societies. That’s personally not something I would like to see either, but if individual information is managed in a way that protects privacy, they could be used to offer greater convenience in society.
AI cameras could be used to immediately locate a child who has got lost, or to apprehend a criminal as soon as they show their face in public. Another use could be being able to see whether a shop is busy to help a person decide whether to go in person or to get delivery.
Spatial data conversion is a technology that will optimize and offer greater convenience in society. It’s a sector that should be developed carefully, but I’m also excited about delving deeper and uncovering new possibilities.
AI is just a tool. Humans use tools to develop and move forward. Just as cars and smartphones have made our lives more convenient, so too will the smart application of AI lead to a prosperous society. I want us to adapt to the changes that will come in the near future, not fear them.
・A conviction that AI will become a key technology like the Internet that changes the way people live.
・Kitade had no experience with AI but felt that even if he failed doing something he didn’t understand, he would still gain knowhow and develop skills.
・Kitade focused on AI cameras because he thought that the company would fade away if it stayed as an AI generalist.
・Image analysis technologies can be used wherever people gather, but it’s difficult to scale a product in a specific sector until you’ve worked out what its fit is.
・It’s chance that determines whether collaboration on an accelerator program will go well.
・The proximity of physical distance and the enthusiasm of those involved help make a collaboration successful.
・Vision of a hybrid of people and technology that is achieved through realspace data conversion and labor saving.