The promise of AI is easily understood by anyone with an imagination, and for 40 years, venture capitalists have been enthusiastically investing in that promise.

However, it’s been significantly harder for founders to turn that investment into sustainable business models. 

Today we are going to look at why that is, and go over what might be a blueprint for startups to create business models around artificial intelligence.

Tatsuo Nakamura founded Valuenex in 2006 with the goal of using artificial intelligence to supplement the work being done by patent attorneys, and their software was instrumental in the resolution of one of Japan’s most famous, and most valuable, lawsuits.  the Blue LED patent case.

We also talk about how to sell to large companies as a small startup, the challenges in trying to make product strategy based on technology, why staying private longer is not always a good thing for startups, and how Valuenex technology accidentally discovered a secret collaboration between Honda and Google.

It’s a great discussion with the founder of one of Japan’s most successful AI companies, and I think you will really enjoy it.

Podcast: Play in new window | Download | Embed

Subscribe: Google Podcasts | Stitcher | RSS | More

Show Notes

  • Why AI can understand patents better than lawyers can
  • Why the market should drive technology rather than the other way around
  • How Valuenex helped resolve one of the biggest patent lawsuits in Japanese history
  • How a new law if forcing change in Japanese universities
  • How Valuenex discovered a secret collaboration between Honda and Google
  • How to create sustainable business models in AI
  • Why quantum computing will both break AI and save AI
  • Why Valuenex IPOed early instead of staying private and growing
  • Some unusual advice about when to do a market entry
  • Why Japanese VC often make market entry difficult

Links from the Founder

Leave a comment

Transcript

Welcome to Disrupting Japan, straight talk from Japan’s most successful entrepreneurs.

I’m Tim Romero and thanks for joining me.

Today, we’re going to be talking about something that’s frankly difficult to talk about on an audio podcast. Tatsuo Nakamura founded Valuenex in 2006 to use Artificial Intelligence and modern visualization techniques to help clients make sense of their patent portfolios and to keep an eye on what the competition is doing. In fact, this technology uncovered some of the core evidence that decided the famous blue LED case. It’s highly effective but highly visual, so let me try to explain it.

Valuenex creates a kind of topographical map that shows companies where in the market, their IP is strong and where it’s weak. This can let them spot new market opportunities or learn what their competition is about to do. It’s all pretty intuitive when you see it, but today, we’ll have to use our imagination as a kind of screen simulation.

Tatsuo and I also talk about Valuenex’s US market entry – well, their two US market entries, actually. We cover what he sees as the best overall strategy for AI startups for them to find their product market fit, and Tatuo explains how he was accidentally able to discover a significant collaboration between two world-famous companies six months before the project was announced.

But you know, Tatsuo tells that story much better than I can, so let’s get right to the interview.

Interview

Tim: I’m sitting here with Tatsuo Nakamura, the CEO and founder of Valuenex. So, thanks for sitting down with me.

Tatsuo Nakamura: Thank you very much.

Tim: Now, Valuenex is a leader in visualization and big data analytics and it’s so hard to talk about visualizations on an audio podcast.

Tatsuo: Yes.

Tim: But we’re going to try. So, what’s the best way to explain? What does Valuenex do?

Tatsuo: Valuenex is a predictive analytics company using big data analytics. So, the main purpose is finding the future situation and making the strategy for our client.

Tim: Okay, and you specialize in intellectual property and patents, right?

Tatsuo: Because patent documentation is very, very good and useful for data source, because there’s many fields in the data and they have a very clean documentation, so it is easy to analyze.

Tim: Patents, especially, are interesting. I’ve heard from a few people in AI that right now, AI can read patents better than humans can because the language is so specialized and unusual that’s it’s really well-suited to AI.

Tatsuo: Yes, so at first, when the people entered the IP field, “Oh, this is confusing because this is not language.” It’s a special words. However, when you apply machine learning technologies, it is easy to understand it because it is a very clear some kind of mechanism for the sentence. So, it is easy to transfer to the machine languages.

Tim: Okay, so once the algorithm goes through hundreds of thousands of patents and uses machine learning and AI to understand the contents, it builds kind of clusters of meaning around it, right?

Tatsuo: Mm-hmm. Can you imagine? It was a 400 documents are there so how many relationships each side?

Tim: Well, it is geometric, right? So, it will quickly get into the tens of millions of…

Tatsuo: Yes, it is huge relationships.

Tim: Right, so a human being cannot possibly understand that from text information.

Tatsuo: Yes, we use a very special person to try reading each by each, so probably, it takes over 10 years.

Tim: Have a bunch of patent attorneys, put them in a room.

Tatsuo: Yeah, but either he or she is in the Japanese major company, IP chains with over 50 people working for the reading for the analytics. After five years, good. So, that, of course, this is the old style. Now, the people are using the analytics tools.

Tim: And so, Valuenex creates the visualization. The output is sort of like an IP heat map, so the users can see where the IP is clustered.

Tatsuo: Basically, it’s three layers. When we think about the next products or next markets based on the technologies, at first, we describe the technology mapping, technology landscaping, and finding the significant technology and connect it to the product, either the competitive product or service, so if there’s no products or services, or this is a good opportunity.

Tim: Okay, so you could use the visualization to say, “Okay, we’ve got a hot cluster of technology in this space, but it looks open in terms of products in the market”?

Tatsuo: Well, most of the cases are market-driven analytics, so the people who want it have a lighter device, then they try to change the materials, but along this side, everybody is trying to make a new idea, so they want to know where is the market?

Tim: Oh, wow, yes! So, I guess it can go in both directions. So, tell me a bit about your customers. How are your customers mainly using this?

Tatsuo: At first, we started, our service is R and Design, so my original company is R&D division and IP division, they have many projects, many ideas, so I guess 99% idea in technology was not released. However, I think that there’s more opportunities and more good services, then they have to use the panoramic view analytics describing, oh, this idea, this technology is very similar and they can be combined.

Tim: So, is that the most common use case where your customer is using it bottom-up?

Tatsuo: Ideally, it is a very good way. However, any company has divisions and they make a wall, secret walls, so they can’t reach into other fields.

Tim: I could see that, yeah, the technology, it seems ideal for just looking and saying, “Hey, you have an opportunity,” but the different divisions in the company force them to have more of a product focus.

Tatsuo: So, if someone thinks about, “Oh, this technology is very good for the battery technologies,” however, they are just materials, so they have to collaborate with electric filed persons. They can’t work that.

Tim: So, for example, let’s say I was the division chief at a big database company and responsible for the automotive industry, could I say, “Let’s look at all of our patents and see what might be suitable for the automotive industry, what blank spaces there are that we could exploit?” Would that be a good use of Valuenex?

Tatsuo: Yes, although a concrete case is a chemical company, Asahi Kasei. At first, they rejected us because they had an analytics genius, but when they used our methodologies, this is good work for the Asahi Kasei president – company president – because the big company’s president has some kind of image for each division’s relationships. So, our or radar chart is almost the same with the president’s idea. So, we can easily point out, oh, this division should be the collaboration or both divisions have white space, so collaborate on the bridge technologies.

Tim: This is really interesting. It sounds like you took the customer’s biggest objection to your product that was bottom-up and we hit all these walls, and you said, “Oh, well, let me explain these walls to you.” That’s great.

Tatsuo: Yes, I established in 2006 this company, so always, the starting point is a negative start.

Tim: Well, actually, I want to drill down deep into the marketing and to the technology in a minute, but before that, let’s step back a bit and talk about you, and when you founded Valuenex, so at Stanford, you and I were talking and you mentioned the importance of the blue LED case, and can you tell me? I think that’s really interesting.

Tatsuo: So, this example case is very famous in Japan and this is a very big event for the society and also me. However, some parts are very confidential, but I can say I was the project leader at these court issues, I belonged to the Nichia Chemical Companies. At the time, 2005, so many people want to know what are these court issues related to this project? So, Nichia Chemical Companies’ IP division is responsible to ask us to use radar chart. Just using this radar is very helpful to explain to the courts. Just two points, one point is the main patent is we call the 404 patent. This is 404 patent has a similar patent with Mr. Akisaki who had issued the patent, advanced to Nakamura Shuji, and the other is NTT’s patent, and NTT’s patent and Mr. Akisaki’s patent is very similar to the 404 patent, but the Japan patent office wasn’t aware of that.

Tim: So, what the visualization showed was that this 404 patent was not as unique as the researcher was claiming, and so he shouldn’t have been awarded the full amount he was asking for.

Tatsuo: Yes, and another point is, Nichia Chemical Companies didn’t use new 404 patent. So, their core area is close to the 404 patent. However, when describing the radar chart, Nichia Chemical’s acquired technology is here, far from the 404 point here. There was no bridge, no relationship for each side, and so you sort of improved the usefulness of the technology or at least the approach in that patent lawsuit, and so were you still teaching at Waseda at that point?

Tatsuo: Yes, yes, yes, Todai and Waseda.

Tim: Was this technology coming out from university research?

Tatsuo: No. Yeah, sometimes, our methodologies are generated in the university. No.

Tim: But in this case, no, it came from outside?

Tatsuo: Yeah, so stakeholders from Waseda University and maybe individuals within venture capitalists of Waseda University.

Tim: Okay, did you always want to start a company and just happened to be teaching until you could make that happen, or…? Because it’s still pretty rare for people to transition from being a university professor to running a company.

Tatsuo: I’m a little rare in Japan from the academic side and the business side, but in Silicon Valley area, there are many startups related to the universities. It is a very popular case.

Tim: I think at Stanford, the professor who doesn’t want to start a startup is very unusual.

Tatsuo: Yeah, it’s a good situation.

Tim: But in Japan, it’s the opposite.

Tatsuo: Yes, in Japan, the professor is just always in universities, so they don’t know the outside of the universities. this is not good for the student.

Tim: I agree. So, before you started teaching, you were at Mitsubishi Research for a long time, so it’s a very interesting and unusual career path in Japan – it’d be very normal in the US, but to go from an organization like Mitsubishi Research to the University of Tokyo, to running a startup, it just sounds like some really big changes. I’m curious, like why, what was your motivation at each time?

Tatsuo: At first, the start of my business life is at the Mitsubishi Research Institute and three years later, I transferred to the University of Tokyo because at this time, the Mitsubishi Research Institute and METI, and University of Tokyo were collaborative, so I worked three years and a half at the University of Tokyo, and after that, I came back to the Mitsubishi Research Institute and worked there for nine years. So, in 2001, I made a business strategy plan but at that time, so it was not a Big Data.

Tim: No, no, it was before cloud computing, before big data, yeah. Okay, so you always kind of had a foot in both worlds; you always had a connection to the business side and to the university side.

Tatsuo: Yes, yes, and now, we’re working with Waseda University.

Tim: Is that something that’s changing in Japan now? 20 years ago, I mean, you would never hear of a university professor on the board of a company or on a startup board. But we’re seeing more and more. Is that something that’s changing, do you think?

Tatsuo: Yeah, I think so. These 10 years, very changing, because universities have to thrive, so yeah, I don’t remember the accurate year, but around 2003, Monkasho changed their policy to supporting the university. Just only 40 universities received budget from the government, so it started the competitiveness in the universities after 2003, so they ordered hiring talented professors from outside the university.

Tim: Ah, okay, so the universities are kind of being forced to change? Well, that’s the only way they’re going to change. Well, just thinking about it, I mean, it would seem that the ideal use for Valuenex would be working with universities to look at their technology and their patent portfolios because there’s a tremendous amount of IT at universities and Japanese universities are terrible at commercializing it. I mean, you must have proposed this to the universities already. Have they been interested in it?

Tatsuo: So, my case is universities are good to stock for the students. It’s a good relationship.

Tim: That is important.

Tatsuo: And good future customers. Yeah, so every October, I have a class and several students come to our office as interns.

Tim: Yeah, I can see why that would be very valuable for both sides, but don’t the universities themselves want to use Valuenex to, for example, look at their own patents and look at their own technology and say, “Hey, we have this great cluster in healthcare,” or “What we’re doing now could really be applied to driverless vehicles.” It seems like such a good fit.

Tatsuo: Probably in 10 years or 20 years, the professors are changing, so the next professor generation is like me.

Tim: We have to wait a whole nother generation? I don’t want to wait that long.

Tatsuo: Yeah, it takes a long time.

Tim: No, but I mean, it seems like what you have is more of a prioritization tool. It’s not like a black box, but it would tell those 20 patent attorneys, “You should look at this cluster.”

Tatsuo: So, our tool is not suitable for the typical IP business, yeah, because our business is so focused on making strategies, finding future situations. Our priority such – of course, we can do that but this is not so different from just searching.

Tim: So, when I saw a mention of like, use Valuenex to find, like an intersection in Google and Honda.

Tatsuo: Ah, yes.

Tim: Can you talk about that because I thought that was a great example.

Tatsuo: Yeah, 2013, when I visited Google’s head office for the discussion about IP issues, I made this sample case, so because Honda Research Institute is close to the hometown head office, so just reason – this is just a reason for why I use Honda and Google. So, I made the Honda – whole patent and the Google whole patent, of course, it’s far from each other, but between each core technologies, this is the same area – this is CCTV Camera Technologies. So, Honda and Google have the same patent here, so I asked Google, “In this case, are you competing with Honda or are you collaborating with Honda?” Nothing. He rejected the comment because –

Tim: So, you knew that was a great question.

Tatsuo: Yeah. So, six months later, they collaborated with each other.

Tim: Okay, so yeah, that sort of heatmap, those clusters told you something was happening.

Tatsuo: Yes.

Tim: Excellent.

Let’s talk a bit about kind of AI business models because AI, artificial intelligence, over the last 40 years – maybe 50, it keeps going in these cycles, right?

Tatsuo: Yes.

Tim: And before, you mentioned that part of year business is in consulting and part of it is in like, platform and tool sales. Can you say about, like, how much of your business is platform and how much of your business is consulting?

Tatsuo: Our goal is providing the solution for the customers, so the three types of methodologies is just providing our tools and second is just consultations, and the third is using the tools and consultations because just two provides just 70% with customer satisfaction, so we have to add more value by the human interest.

So, this is kind of a limitation, so 70%.

Tim: Of course, everyone wants to sell software because that scales the best.

Tatsuo: Yeah.

Tim: So, the consulting you do, is it mostly kind of like training or is it data cleansing? What kind of consulting do your customers need?

Tatsuo: So, our consultations, it’s making inside, so how to extract the inside from the analytic result.

Tim: Okay, so helping your customers make sense of the results. Telling them what it means.

Tatsuo: Yes.

Tim: Actually, it’s been interesting – just in the last, say 4 or 5 years, the AI industry is changing pretty quickly, so like four years ago, there was a lot of sort of general AI companies, companies saying, “Just give us your data and we’ll show you the insights,” but it seems these days, the AI companies that are successful all tend to have very narrow specialty.

Tatsuo: Yes.

Tim: So, whether it’s IP, whether it’s like fault detection, do you think that’s the best strategy for AI companies? Do you think they have to focus on a small…?

Tatsuo: Yeah. So, the very famous AI companies focus on manufacturing improvement. They released one of the success stories, so they can reduce costs because manufacturing line, improving, but this is just focused on the special lines, so they can’t provide these algorithms to another line. Which is better? Quality is not good, but they’re not using many persons, or the specific customizing and the very expensive algorithm but these are very good qualitys.

So, both are different.

Tim: Yeah, it seems like the dream of AI and always the promise of AI is the black box solution: just the answers come out, but it sounds like in your case, you were saying like your customers need help understanding what the results mean.

Tatsuo: Yeah, so X axis is the quantity and the Y axis is quality, our service is some kind of here, but just visualization services are not so high-quality, but the huge market. Our service is not so many markets but high quality, high services.

Tim: Right, so selling the services and the understanding that go along with the AI?

Tatsuo: Yes. So, the AI is improving rapidly,  and finally, we are moving here.

Tim: Okay, so you think over time, as AI improves, you will be able to do less and less consulting and let the platform do more and more of the work.

Tatsuo: Yes. So, our methodology is using the batch file technology, so this is not real-time, so you have to wait several minutes or sometimes, several hours.

Tim: Well, I think for like IP analysis, that’s no problem, I’d imagine.

Tatsuo: Yeah, of course, of course. So, 100,000 documents should take over ten years, but in just one hour we do it. But most people that are familiar with search engines just point to second coming results, so we have to think about how to reduce the time.

Tim: What do you see as the biggest challenge in making that happen? Is it the AI algorithms have to get better? Is it that you have to understand the domain better? What has to change?

Tatsuo: I think this is optimization methodologies. This is not AI. Optimization methodology is realized by quantum computing. So now, people are focused on AI, but I believe the next stage should be using quantum computing.

Tim: Well, quantum computing, it’s hard to know. We always seem so close to it, but we always seem to be like, two or three years away from quantum computing.

Tatsuo: Yeah. Yes, so five years ago, all scientists said quantum computing will be realized after 2030, but D-web, other quantum computing companies started small services from next year.

Tim: Let me ask you kind of the different questions. So, you went public last year and there’s been a real trend in the last few years among Japanese startups to stay private longer and delay the IPO. So, what was Valuenex’s motivation for IPO-ing as opposed to staying private and keeping that flexibility?

Tatsuo: Well, the purpose is very simple, so we want our methodology to expand worldwide, so how to realize worldwide activities.

Tim: Was going public necessary for funding reasons or like, reputation reasons?

Tatsuo: Both yes. So, funding reasons is for starting next generation’s development and reputationally, it’s hiring the good quality person.

Tim: Actually, yeah, let’s talk a bit about your global expansion and your market entry strategy because that is something that all high-quality Japanese startups are thinking about going global now. You began your US market entry 2014, was it?

Tatsuo: Yeah, 2014. So, at first, when I started by 2006 in Tokyo, next year, in 2007, I opened the San Jose office.

Tim: So, you opened a San Jose office the next year?

Tatsuo: Yes, because when I established my company, these methodologies should be expanding worldwide from the start.

Tim: That’s really fast to go global. So, again, was your goal to establish credibility by having a US office? Was it to like, try to reach US talent for programming and ideas, or was it sales? What was your goal that early?

Tatsuo: Yeah, so at the time, just starting the startup and just entered the US side, I don’t know the management. Of course, I don’t know the relations – what is an attorney? No, no, no, no, everything you know. So, I supported from the San Jose incubation office. Fortunately, I had success with the fundraising twice at that time, so we had some kind of huge money at the time. However, in 2009, the Lehman shock was all over, so our stakeholders ordered me to go back to Tokyo and close the San Jose office, 2009.

Tim: Oh, so 2014 was actually going back to the US?

Tatsuo: Yeah. So, it is a good story and bad stories. Bad story is in 2009, at the time, there was no big data, no AI analytics tools in the US side, so when I reentered in 2014 the US side, of course, there were many AI analytics and big data companies. So, we had to catch up with these companies. The good story is, during the developing our tools in only the Japanese market, we developed this in an original way. This is a different way for big data analytics because the US’ and worldwide’s big data analytics is expanding this side.

Tim: So, trying to be more – yeah, everyone wants to be the platform –

Tatsuo: Yeah, platform.

Tim: Without specific expertise.

Tatsuo: Yes, not qualities, but our side, our methodology is keeping the quality, so this is the reason why we are now competing in the US side.

Tim: So, it goes back to those clusters again. Everyone else was clustered around the platform business and you were providing very specific technology, okay, but yeah, opening a US office one year after you start the company, is that something you would recommend to other startups to do?

Tatsuo: No. Yeah, in fact, at the time, very hard-hearted for me. Yeah, so funny story, so me and sometimes, my member driving the car by rent a car, sometimes, enter the tram line. The police called us and stopped us, “Oh, you are junk driver.” Oh gosh, there’s many troubles.

Tim: So, these days, do you spend most of your time in Tokyo or most of your time in San Francisco?

Tatsuo: Now, after 2014, half of the year is staying in Tokyo and half is staying in the US.

Tim: Alright. So, you were mentioning that the year after you start your company is far too early to do a US market entry. So, what would your advice be to Japanese startups? When should they start thinking about doing market entry?

Tatsuo: Typical idea for the Japanese startup companies, at first, they’re making the market share in Japan and after, they want to enter the US, this is not good for going to the US side. Yeah, because market culture and the UI/UX is always different between the US and Japan, so if they want global market, they have to think of going to – at first, US or global.

Tim: So, when do they know it’s time? So, is it like, after you have product market fit with your Japanese customers, is it after you raise a certain amount of money, is it just as soon as possible? When would you advice like, the right time?

Tatsuo: So, it depends on the field, so fintech and big data analytics has a four-year difference between the worldwide market and the Japan market, and the type is big difference. However, biotech, some kind of material science, has a big advantage in the Japan side. So, they have to keep their technology with the ideas of IP and to consider how to enter the US market.

Tim: Alright. A lot of startup founders have told me that the VCs give them a lot of pressure not to do market entry until after the IPO because a lot of Japanese VCs, they’re focused on that IPO, and it’s okay, but they’re saying, “No. Market entry, that sounds risky and expensive and why don’t you just wait?” but you started your market entry twice, both times before the IPO. Did your investors complain about that or were they supportive of that?

Tatsuo: Very fortunately, the lead investor is very supportive of me, so over ten years holding the stock share and after IPO, they still didn’t release their stock.

Tim: Wow!

Tatsuo: This is very strange in Japan.

Tim: It’s like ideal investor.

Tatsuo: Yeah.

Tim: Just keep it forever.

Tatsuo: Yeah, they want to keep it forever, and also, they’re looking for a good collaborator for the next stage stakeholders. So, these are important things, so how to collaborate with a startup founder and venture capitalists.

Tim: And, as part of that market entry into the US, were your first US customers subsidiaries of your Japanese customers or did you have to develop a customer base from scratch?

Tatsuo: When we entered the US side, soon, I was aware that we had to change the UI/UX and the activities, so we made a contract with Silicon Valley developers and changed most of the parts of the UI/UX.

Tim: Okay, so for the business side, it was really starting from zero again?

Tatsuo: Yeah, yeah.

Tim: Alright, okay.

Well, listen, Tatsuo, before we wrap up, I want to ask you what I call my “Magic Wand” question and that is, if you I gave you a magic wand and I told you that you could change one thing about Japan – anything at all – the education system, the way people think about risk, the legal system – anything at all to make it better for startups and innovation in Japan, what would you change

Tatsuo: Yeah, I want to change Japanese people’s mind, because in the long history of Japan, Japanese people, basically, they want to go abroad, but now, these 20 years, they’re keeping inside; they don’t want to go abroad.

Tim: So, why is it important that Japanese people go abroad?

Tatsuo: Because this is not only for business. Private is okay, so communicating with other cultures, persons, people are aware, “Oh, this is different,” or “This is another way.”

Tim: Is it the advantage that they see new ideas or just gets more flexible thinking?

Tatsuo: So Japanese think about the former ways. Currently, people can easily find just searching, but they have to think before searching, what should we do? How should you do that? So using another culture, another people, another way, diversity is very important for making the new insights.

Tim: Okay, so it just makes people aware that there are different ways of thinking and you have to approach problems differently.

Tatsuo: Yeah.

Tim: Yeah, I think there’s a general agreement that kind of in principle, diversity is good, but as you mentioned, Japan is in an interesting situation now where because of the internet, in some ways, it’s easier than ever for people to be exposed to new ideas, but as you mentioned, travel’s actually decreasing and what do you think is going on there? Do you think people are getting more closed or more open to outside? What do you think is happening?

Tatsuo: Yeah, so after the collapse of the bubble of 1991, people became very conservative. They don’t want to challenge everything.

Tim: Okay.

Tatsuo: So, we are the last generation for challenging. So, the problem is the 30-years generation and and 40-years generation can’t be in our situation. Why do that? There are many risks. But we want to do that. That reason is very simple.

Tim: Do you think that’s changing with the younger generation with the Japanese who are, say college students now?

Tatsuo: Yeah, yes. Until four years ago, my college students want to be the bureaucrat or joining major big enterprise companies, but recently, little changes. Some part of the students study the startup company during student life. I expected this situation increasing.

Tim: It’s progress.

Tatsuo: Yeah.

Tim: Ah. Okay, well, listen, Tatsuo, thank you so much for sitting down with me.

Tatsuo: Yeah, thank you very much. I was very excited about this discussion.

Outtro

And, we’re back.

You know, I found Tatsuo’s comments about the challenges of selling Valuenex into large companies to be a really good example of why it’s so hard to sell innovation in general. Valuenex can be used to spot new market opportunities, to open new lines of business. You think it would be easy to sell, but corporate sales doesn’t work like that. When you’re selling to a large enterprise, you’re not being paid to solve that company’s problems; you’re being paid to solve the problems or whatever specific group or division is signing that contract. If the problem falls outside of their responsibility, then it’s not their problem and they probably don’t want to use their budget to solve it.

The challenge here is that disruptive innovation or really, just innovation in general usually requires the involvement of several different groups or sometimes, paying attention to a problem that no one has specific responsibility for.

In the story Tatsuo told about Asahi Kasei, he had to go all the way up to the CEO before he found someone who understood the importance of cross division corporation, and that is a problem. If you always have to go to the top, it’s hard for real innovation to get traction.

Also, Tatsuo gave some really interesting advice about market entry and AI market strategy in general. Now, going into global markets as early and as hard as Valuenex is the opposite advice that is usually given to startups, but it’s clearly worked for them. They were one of the handful of targeted AI companies with real domain expertise in a market flooded with generic AI platforms, so they got the attention they deserve, but it’s easy to understand the appeal of blackbox AI.

I mean, the blackbox AI where you simply feed in data and it gives you the answer has been a staple of both science fiction and investor interest since at least the late 60s. If such a device could be invented, the founder of the company would undoubtedly become the world’s richest person, so I get the appeal, but is the blackbox AI possible? I mean, is it even theoretically possible?

I’ve been to a few generations of the AI hype cycle and I’ve noticed that every time, there’s a lot of head waving away of the real human expertise and the real human problem-solving that goes into creating an AI answer. It’s not just the data collection and data cleansing, but figuring out exactly what the right question is and the precise way to state that question, and of course, how to interpret the AI’s output, and in most real-world situations, it seems that the bulk of the value is in determining exactly what question we need to be asking, and that question is almost never the question that we start out asking.

AI is an incredibly useful tool, of course, but at the moment, it looks like we humans should count on having to solve our own problems.

If you want to talk about AI or selling technology to large enterprises, Tatsuo and I would love to hear from you. So, come by DisruptingJapan.com/show151 and let’s talk about it. If you leave a comment, I guarantee you that Tatsuo or I, or maybe both will respond, and hey, if you get the chance, check us out on LinkedIn or Facebook, but even better, if you like the show, tell people about it. Disrupting Japan is grown not by social media marketing or advertising but because listeners like you enjoy it and they tell their friends about it.

But most of all, thanks for listening and thank you for letting people interested in Japanese startups know about the show.

I’m Tim Romero and thanks for listening to Disrupting Japan.