[Interview] A leader in advancing digital transformation of infrastructure inspections using cutting-edge technology

Convergence of Applications, Platforms and Data

Nishimura:Your company provides solutions tailored to the issues each customer faces, based on your core technology, SENSYN CORE. However, this explanation alone may make it difficult to imagine what exactly it is, so could you tell us more about the appeal of SENSYN CORE and how you apply it to your customers?

Kitamura:Please see the architecture here. In our business, we control smart devices and use the data obtained from them to gain insights that detect abnormalities and changes. To achieve this, we provide business applications that are easy to operate even for end users with low IT literacy and can be used for a long time. In addition, we own the SENSYN CORE Platform, which makes it possible to extract and utilize data with high reproducibility. We develop optimal solutions by combining applications, platforms, and data.

Nishimura:Could you please explain in more detail what the SENSYN CORE Platform you just mentioned is?

Kitamura:The SENSYN CORE Platform has three main functions. The first is SENSYN Edge, an edge technology for controlling smart devices like the ones I mentioned earlier. When collecting images, videos, temperature, and other data in harsh or high-risk locations for humans, it is necessary to automatically operate drones, various robots, and tools known as smart devices. Our SENSYN Edge makes this possible.

The second is a cloud service called SENSYN Data. When the data actually collected by smart devices is uploaded to our cloud service, the images and map data are linked, 2D reconstruction is performed, and the data is organized chronologically with meta information. When these are accumulated, they become big data and can be used as statistical information.

The third function, SENSYN AI, analyzes this data. This makes it possible to automatically detect abnormalities and changes such as rust and cracks, and is useful for preventive and predictive maintenance. Ultimately, it reduces unexpected repairs and abnormalities, and supports management decision-making such as planned repairs. At the same time, SENSYN AI is also used as the control technology for SENSYN Edge, and we position AI as a core technology.

In this way, we provide solutions by utilizing edge technology, data technology, and AI.

We want to be partners in solving problems together

Nishimura:Could you tell us about your company's strengths and what you think makes it superior to other companies in the face of so many competitors? Of course, I think you are excellent overall, but I would like to hear about any particularly unique aspects.

Kitamura:One of our strengths is that we place great importance on our clients' business processes and data flows. These are broadly divided into three steps. The first is to obtain the data. The second is to analyze and interpret that data. And the third is to utilize that data to have an impact on actual business operations and society.

Manufacturers selling devices such as robots and drones, and companies specializing in that field, tend to stick to the first step of acquiring data. With the intention of selling products, they tend to take the approach of "It's safe if it works in place of people," or "Let's buy it and try it out first." However, we do not take such an approach. First, we identify the areas that will have an impact on business, and then we hypothesize what issues there are and what the impact will be when those issues are resolved, quantitatively and qualitatively. This is because we believe that this method is meaningful both in business and society. To do this, we need to identify the source of the data, such as what insights are needed and where the data to extract those insights comes from.

If it is best to collect data from the air, we have the option of using a drone. However, drones have the risk of crashing during flight, and we have actually had several crashes during testing. Therefore, we use ground-based robots in sites where the risk cannot be tolerated. We also use network cameras and smartphones in cases where high-tech equipment is not required and a simpler method is appropriate. In this way, we are not tied to any specific hardware and offer flexible options. We can use the best technology available from around the world, and in some cases we believe that human data collection is fine.

Our goal is to make an impact on business and ultimately contribute to society. There are not many companies that handle everything from data acquisition to analysis and utilization, but our greatest strength is that we can provide this process in one stop. This has been highly evaluated, and our solutions have been adopted in many fields.

Nishimura:I understand very well. From what you've said, it seems like you're taking a backcasting approach rather than a forecasting approach. In other words, you first thoroughly discuss and determine what you ultimately want to achieve, and then think about how to obtain the data to get there. In that sense, the ability to listen carefully to your customers' voices and identify what the issues are and the true questions, in other words, consulting ability, is extremely important.

Kitamura:This is a very important point, but you should not directly ask the customer, "What is your problem?" This may be done in general solution sales, but if you do this, by the time the customer answers, "This is the problem," there is a chance that they have already told the same story to other companies.

Nishimura:I see. In other words, the problem becomes apparent at that point.

Kitamura:That's right. If the problem is obvious, customers will think that any company will do as long as the price is low and the functions are extensive.

Nishimura:That's certainly true.

Kitamura:We cannot be just another vendor. Our goal is to be a partner that works with our customers to find problems. To do this, we first propose a hypothesis. For example, we suggest, "If you are in this business, do you think you have these kinds of problems or pain points?" The customer may empathize and say, "Yes, that may be true," or the conversation may expand, leading to the idea that "It's not ours, but the next department may have such a problem." When the problem is shared in this way, when the question arises of "So how can we solve it?", our track record and evidence to date can be useful as reference material. We do not simply make proposals, but we place great importance on being recognized as a partner that works with our customers to identify problems and solve them together.

Nishimura:So you work alongside your customers to identify issues and problems.

Kitamura:That's right. If we force our products on people, we cannot become true partners.

Nishimura:In that sense, how do you usually make that first connection?

Kitamura:There's been a lot of word-of-mouth lately, which is what makes me happiest.

Nishimura:Indeed, that's awesome.

Kitamura:We are increasingly receiving consultations from other departments of our clients, such as "I heard you're doing this, what do you think?" and from other companies in the same industry, such as "I've heard rumors, but I'd like to try it out at our company." Information is surprisingly often shared, especially in large-scale heavy industries. We visit many companies, and word of mouth is still the most successful.

Nishimura:After all, pulling is more effective than pushing.

Promoting the "democratization" of tools

Kitamura:Until now, we have often been thought of as a "drone company," but that is because drones are extremely capable and a convenient tool for acquiring data. However, there are many things that drones alone cannot do, so we have been steadily increasing the number of other means and automatically controlling these tools on our platform, promoting "democratization" to the point where anyone can use them safely and securely, even without skills.

As a result, we have been collecting huge amounts of data, and we are now moving into the phase of utilizing that data to create impact. We are particularly focusing on the field of AI. We already had an AI team, but it has only been in the last one or two years that it has really started to function.

To give a concrete example of the impact it is making, let me give an example of the inspection of wind power blades. The level of cracks and rust is judged, and if it is level 2, it is fine to continue monitoring, but if it is level 4 or above, it is decided to move to the inspection process even if it is high or in a dangerous location, and only then will a person be able to directly access the tower using a rope. If repairs are actually necessary, the parts will be ordered and repairs will be encouraged to be made by a partner company. AI is also being used to inspect a variety of other infrastructure, such as cracks in building wall tiles and concrete, and detection of loose bolts in power facility wires and towers.

We also handle automatic reading of pressure gauges in plants and, more uniquely, spark detection in steelworks. In steelworks, safety inspectors are always present to prevent fires caused by sparks that occur when processing iron, but as they are human, they can miss things. Therefore, we have created a system that issues an alert if a spark flies beyond a certain threshold. Furthermore, by recording several tens of seconds before and after the alert, evidence that can be used for training and review can be left.

In this way, we basically start with equipment inspection and maintenance work, but there are many risks and analog areas surrounding it. We listen closely to the risks and issues our customers face, and if we can solve them purely with software, rather than just using drones or robots, we implement that and use the data to make an impact on our customers' operations. That's the direction we're aiming for.

Nishimura:In other words, equipment inspections are just the beginning. They serve as a starting point to get closer to customers, solve their problems, and help them advance digital transformation.

Coexistence between AI and humans is the direction we should aim for

Nishimura:This is a simple question, but could you tell me about the accuracy of information collected by drones and other digital and smart devices? I think everyone is curious about how accurate it is. If it is possible to a certain extent to eliminate the need for human eyes, that would be ideal, and human eyes are still necessary, but can it be judged basically by AI alone? How advanced is the current state of image diagnosis by smart devices?

Kitamura:AI has not yet caught up with human eyes. Humans have a very high ability to use their five senses, such as smell. AI and robots are focused on one thing, so they cannot beat humans in that area. However, AI and robots are overwhelmingly superior in terms of durability and the ability to continue performing the same task with the same precision.

Customers often ask us, "Wouldn't it be more accurate if a human did it?" We say, "If it's more accurate than a human, we'll consider introducing it." But in reality, it's not true that humans can always maintain that high level of accuracy. They may miss something, and as they get tired, their eyesight may become blurred. Furthermore, accuracy can change as people get older. When you think about it that way, it's AI that has the advantage on average.

Nishimura:There may be moments when human eyes are superior, but in order to consistently deliver stable performance, introducing these types of solutions is essential in Japan, where the population is aging and the birthrate is declining.

Kitamura:That's right. It's been said for a long time that "intuition, experience, and courage (KKD)" is important, but there's no point in comparing it to AI. The goal is to have AI take on that role and standardize it as much as possible. So the argument that accuracy must be close to 100% is not very effective. There are cases where it's better to be able to continue working with the same accuracy, even if it's 70%. It's important to have such discussions. The question of "what percentage of accuracy" does not capture the essence. What's important is that machines and humans coexist.

Nishimura:I see. It's true that when you pursue 100%, there are times when you can't see through it. The ideal situation would be one where we correctly recognize the limitations and problems of humans, where machines and humans coexist, and where machines can work at the same performance 24 hours a day, and where the five senses of humans complement each other. So it's important to cooperate, not see things as conflicts.

Kitamura:That's right. I often say that robots, AI, and humans should work together. For example, there are thousands of steel towers for power facilities. If we could use AI to determine, "These 10 towers are dangerous and need to be closely watched by humans," productivity would improve dramatically. Ideally, AI would take on this screening role.

Nishimura:That is a very important message. It is true that we tend to fall into the trap of thinking that machines will do everything, which is an all-or-nothing argument.

Kitamura:Yes, I believe that in this industry, and in other technologies as well, there is more value in humans and machines working together.

Nishimura:You mentioned durability as one of AI's strengths earlier. Can you elaborate on that?

Kitamura:Humans get tired, but AI doesn't. For example, when a drone or robot takes a picture of an object, hundreds, sometimes thousands, of images are collected at once. It is a lot of work for a human to check each image, and no one wants to do it. That's why the processing can't keep up unless it's automated. On the other hand, a human may be able to intuitively decide that "this looks OK" or "this looks strange," but if a human analyzes the data, they will have to check every image in detail, which increases the workload. That's why we need a system that automatically exposes abnormalities using AI. Doing so may make up for parts that humans miss, and it will be possible to detect abnormalities that humans could not find.

Nishimura:Even if data collection by drones were automated, it would be unrealistic for humans to have to review thousands of pieces of data after the analysis. That's why combining it with AI makes it possible to put the data to good use.

Kitamura:That's right. There are many companies that collect data but leave it without processing it. Even if you collect data using tools, if it is simply saved in large quantities on a hard disk, customers cannot use it for inspections.

Supporting the transition to CBM for a sustainable future

Kitamura:So the first thing we did was to create a system that would automatically associate the location of the data taken with the collected data in order to give meaning to it. We linked it to a map so that it would be clear where the data was taken. This would allow data corresponding to a specific location to be visually understood. We then automatically classified hundreds or thousands of pieces of data for each location, enabling us to use AI to find anomalies.

When this becomes big data, it becomes important to find trends. For example, predictions can be made such as "Maybe it's okay to reduce the frequency of inspections for this part" or "This machine is about to break down." It is also possible to take planned action such as "We have the time and workload to repair it now, so let's do it in advance." If this happens, the way work and maintenance are done itself will change dramatically, and tasks that previously required hundreds of people may now be done by just a few dozen. This could potentially combat the problem of labor shortages.

Currently, there is a demand to change from TBM (Time Based Maintenance) to CBM (Condition Based Maintenance). TBM is maintenance performed at regular, set times. However, this method is running out of resources. There is too much aging equipment, and regular maintenance alone is not enough to keep up with it. As a result, maintenance is not done properly, and the risk of equipment failure or accidents occurring during disasters increases.

On the other hand, CBM is an approach that grasps the condition of equipment in real time, detects signs that problems are likely to occur, and takes measures early. Dangerous areas can be identified in advance and responded to in a planned manner. I believe that by shifting to this type of condition-based maintenance, we can build a more sustainable future.

Nishimura:The switch from TBM to CBM is a pretty big topic.

Kitamura:Yes, that's right. They say that it won't be sustainable if we don't do that.

Nishimura:Your company is taking a leading role in addressing this issue.

Kitamura:Yes. Just by talking to our customers, it becomes clear that the on-site environment is extremely harsh.

Nishimura:That's obviously harsh.

Kitamura:That's right. There are many so-called "3K" (tough, dirty, dangerous) jobs, so if people could choose their work, few people would willingly enter this industry. If there is a job that pays more and is not 3K, they will go for that. That's why I think it is important to replace these harsh and risky jobs with new technology. By doing so, new talent who want to work using this technology may appear. In fact, efforts to update infrastructure and legacy fields using robots and AI are very attractive to science and engineering talent. Increasing the number of such talents may also lead to a different kind of elimination of the labor shortage. I would like to work on this in the hope of achieving such secondary effects.

Nishimura:When I looked at your website, I saw the news that you had started a drone academy. I got the impression that you would like to expand the scope of your business by increasing the number of drone users.

Kitamura:In fact, we would like to incorporate the content that I have just mentioned into the academy. There is a reason why we named it "SENSYN ROBOTICS ACADEMY" instead of simply "Drone Academy". As the first step, we are focusing on drones to comply with the national qualification system, but we would like to make it a place to share our know-how and experience, including the successes and failures that we have cultivated so far. Until now, we have not provided it in the form of an academy, but in fact, we conduct a lot of training and education for customers throughout the year. We started this initiative because we thought that by expanding it more widely, we could further increase social value. In the future, we would like to teach how to use drones in various fields, such as AI and other robotics technologies, and provide the necessary training.

Nishimura:I see. So you want to increase the number of people who share the same aspirations and perspectives.

Challenges and possibilities for overseas expansion

Nishimura:Could you tell us about the challenges and strategies you face in increasing the number of clients in the future? You are already working on expanding overseas.

Kitamura:That's right. One of the challenges when working overseas is the huge difference in regulations between countries. Naturally, laws differ from country to country, and often rules and customs are not clearly stated in the law.

Nishimura:It's true that cultures and customs vary from country to country.

Kitamura:That's right. If we handle such things improperly, there is a risk that they will be mistaken for terrorist acts and we will be arrested. Infrastructure is a country's asset, so there is a big risk in expanding overseas alone. After all, we need partners who are familiar with local conditions. Finding such partners is our challenge and also our medium- to long-term strategy.

Nishimura:What roles do you expect your partners to play in particular? For example, in what areas do you expect them to develop sales channels, find customers, secure your supply chain, etc.?

Kitamura:Ideally, we would like to work directly with local customers. Just like in Japan, overseas companies face challenges such as risk reduction and labor shortages, as well as technology standardization and a lack of engineering resources, so we would like to promote digitalization that can meet those needs. That is one of our major goals.

However, this alone exposes us to risks specific to overseas markets, so it is also important to partner with companies that supply facilities and equipment locally. Such companies are familiar with local rules and customs and have been involved in maintenance and other activities for many years, so we believe that utilizing such partnerships will help us avoid risks.

Nishimura:That's exactly right. The Tokyo Consortium is also taking a similar strategy, approaching Japanese companies that are already doing business overseas and aiming to expand locally. Also, regarding the long lead time that is often cited as an issue when expanding overseas, this isn't a problem that is limited to overseas markets, is it?

Kitamura:That's right. Especially in the heavy and large-scale industries, projects progress slowly. It requires risk management and business coordination for the client, and approval and safety assurance for all involved. I understand that, but to be honest, sometimes it feels too slow. That's why we actively issue press releases every month to accelerate the process and deepen understanding.

Nishimura:It takes courage to issue a press release for something that is still a new initiative.

Kitamura:That's right. There is a risk that others will copy us, but more than that, we take pride in being market makers. We believe that the most important thing is to gather our peers and get our customers to empathize with us, so we take the initiative in disseminating information.

Nishimura:This point ties in with what we said earlier about the academy.

Kitamura:I agree.

Nishimura:The role of market maker felt right to me.

Kitamura:You have to create the market yourself.

Nishimura:Do you feel there are any differences in the way discussions are conducted in Japan and overseas? Your company operates both domestically and overseas, so I would like to hear about any differences you feel.

Kitamura:At first glance, don't you get the impression that decisions are made faster overseas?

Nishimura:Yes, there are. It seems like it would be easy to say "yes" to someone.

Kitamura:But in reality, it's not that fast (laughs). In fact, it can be slow. It's not that different from Japan.

Nishimura:I see. I had the impression that Japan was a lot slower.

Kitamura:For example, if it were something like introducing chat GPT, we might make an immediate decision, but if it's about protecting infrastructure, we have to be careful. Failure is not an option, and if an accident occurs, we will be held responsible.

Nishimura:What are the points in the process that seem to move more slowly? Are there any phases after door-knocking where things don't move forward?

Kitamura:That's right. It's a challenge we face as well, but for example, when it comes to solutions that use robots, a big problem is which hardware to use. We can deal with it with software, so it's fine if there is a compatible device, but there are issues such as whether there are devices that can be used in the country and whether the protocol will work. The key is whether such a combination of hardware and software can be introduced locally, and it takes a considerable amount of time to make such adjustments. The communication bandwidth also comes into play.

Nishimura:I see. So device settings are a big factor.

Kitamura:Yes. Plus, we need to find a player who can do the maintenance locally.

Nishimura:You would also need someone to take care of maintenance. That would certainly take a lot of time.

Kitamura:That's right. It would be fine if everything could be done using just a computer or smartphone, but issues of economic security and infrastructure are also involved.

Supporting decarbonization efforts

Nishimura:We have been supporting you as the Tokyo Consortium up until now, but what are your honest thoughts on this?

Kitamura:We worked together on a simulation of decarbonization. They calculated whether our solution would really be effective in deoxidizing, and if so, how much impact it would have, based on the standards of the Ministry of the Environment. As a result, we were able to confirm that it would be effective after all. Of course, that alone will not be enough to get our solution adopted, but I think it is significant that it has a positive impact as a secondary effect. In the future, we will likely have to include our CO2 reduction efforts in our securities reports. I also think that this will be a factor in us receiving a positive evaluation from investors from the perspective of ESG management. In that sense, I am very grateful that we now have the basic data.

Nishimura:I'm glad to hear you say that.

Kitamura:To be honest, that kind of work is a pain. It's mundane work.

Nishimura:Well, no, that's our job.

Kitamura:But it would be really helpful to have someone who is good at that kind of thing, someone who can combine various data and draw a conclusion. I'm sure it's possible to do it by brute force, but there aren't many people who are good at that kind of thing or who like to do it.

Nishimura:It would have been the best scenario if the investment had gone through for that project, but things don't always work out that way.

Kitamura:But I was still impressed with the results. It had such a positive impact. But it doesn't mean they'll adopt our solution. (Laughs)

Nishimura:Well, there are side effects to that (laughs).

Be the first player to be consulted

Nishimura:As you take on your role as a market maker, how do you want to develop Sensyn Robotics in the future? 

Kitamura:First of all, and this is something I've been communicating to the entire company, we want to be the first company that people think of and consult with when it comes to using drones and robots to maintain and protect infrastructure.

Our value lies in not simply proposing solutions when we are consulted, but as I mentioned earlier, identifying the real issues, presenting the optimal solution as quickly as possible, and supporting the implementation. To achieve this, we are always trying to incorporate cutting-edge technology, although there are still some areas that are lacking. Recently, we have been incorporating it into our work, and since it has a high affinity with technologies such as LLM (large-scale language models) and generative AI, a major pillar of our strategy is to quickly adopt these technologies and provide their effects to our customers.

Looking at our relationships with our customers, we are already able to reach out to owner-occupied companies with many facilities, but we would like to spread the wave of digitalization to the affiliated companies in the surrounding areas. We are incorporating this as one of our growth strategies.

Nishimura:What exactly are affiliated companies?

Kitamura:For example, the owners of electric power companies, steel mills, and oil plants are the companies that own the facilities, but the companies that build them are engineering and construction companies. These companies order a lot of materials and have many different contractors coming in and out.

Nishimura:So it will be expanded to all ecosystem players.

Kitamura:Our customer base is very broad. Until now, we have targeted the top of the pyramid, but if we can influence the entire pyramid, our business will grow and we will be able to have a positive impact on the entire industry. We want to make this a reality.

Nishimura:So you start with the facility owner and provide solutions to the entire supply chain.

Kitamura:Yes, that is true. I think that is the core of our business. We are expanding overseas, but we have a strong desire to fundamentally solve social issues in Japan, so we want to continue moving forward without wavering from that point.

Nishimura:Thank you very much.