Episode 13

AI Simplified

with Kai-Shing Tao

Kai-Shing Tao

Co-Founder & CEO

In this episode, Tee sits down with Kai-Shing Tao, Chairman & CEO of Remark Holdings to talk about the power of AI and how the company is providing practical and accessible solutions to their clients that solve problems and reduce risk. 

In this episode, Tee sits down with Kai-Shing Tao, Chairman & CEO of Remark Holdings to talk about the power of AI and how the company is providing practical and accessible solutions to their clients that solve problems and reduce risk. 

This wide-ranging conversation includes topics such as: 

  • How Remark got its start? (2:02)
  • Remark’s AI Technologies (4:33)
  • Companies that use Remark’s Technology (8:38)
  • How COVID has Opened Doors for Expansion (10:50)
  • Remark’s Growth in the U.S. Market (12:53)
  • What does the future look like in the space? (17:32)

Remark Holdings primarily focuses on the development and deployment of artificial-intelligence-based solutions for businesses and software developers in many industries.  Additionally, the company owns and operates digital media properties that deliver relevant, dynamic content.

Automating The Chain bridges the learning gap between business executives and their technical counterparts. Each episode we learn from CTOs and experts in industrial automation as they explain their technology in an accessible way. For more information, or to subscribe, please visit https://www.automatingthechain.com/.

Share this Episode:

Share on linkedin
Share on twitter
Share on facebook
Share on email

Transcription

Tee Ganbold  0:09  

Welcome to Automating The Chain, the weekly podcast and webinar specifically engineered to support and educate executives as they explore the potential of industrial automation. Each week we sit down with an executive leader or their technical counterpart of an international organization to discuss how they plan to leverage industrial automation to advance their business, who also have startups focused on automating the supply chain, explain the technology in an accessible way. Experts in the field will color in historical and current case studies.

 

Without further ado, let’s get into the show.

 

Hello! How are you, Shing?

 

Kai-Shing Tao  0:56  

Good! Doing well. And you?

 

Tee Ganbold  0:58  

Good, thanks. Are you in Las Vegas? Are you in LA right now? Where are you in the world?

 

Kai-Shing Tao  1:06  

I’m in sunny LA and fortunately, even though there’s a bit of a shutdown over here, the weather’s great, so no complaints.

 

Tee Ganbold  1:14  

Good. You guys are so blessed with the weather. Well, I’m so grateful for your time, and I just want to introduce you as the Founder and the Chief Executive Officer of Remark Holdings. And you can introduce the company in a little bit more detail but, not only have you decided to start this company, but you’re really an investor in retail and technology companies. And I guess, I’d love to ask, how did you go from investing in a lot of businesses and then decide, “I’m going to actually focus on building Remark?”

 

Kai-Shing Tao  1:55  

Yeah, I started off as an investor in a bunch of different businesses that Remark was involved in, and then, as we kind of put all the different pieces of the puzzle together, we had the idea of basically data scraping technology (this was in 2013-2014), and it worked really well. We didn’t know what we were going to do with it yet, but it was just a way of testing to see what was out there. We call that the Data Intelligence Platform. And then as we began to read a lot of the white papers that were being written about AI and how, you know, data is needed to— We all say, “Data’s the new oil.” You know, how you need the data to be able to train the algorithms. The cost to create these algorithms and to kind of power overall AI computing had come down to be achievable for a company of our size, and we certainly are not one of the big boys or anything like that. So we took it and kind of ran with it and, five years later, we’ve won a number of pretty famous, prominent computer vision awards and then we’ve been able to do it the last couple of years, year after year. Then most importantly, we’ve been able to take that technology and commercialize it. When you’re public, people care— not about long-term visions, per se, as how you’re going to help them. And so we’ve been fortunate that we’ve been able to test with a number of large companies that believed in AI, created a budget for AI, and now are incorporating our AI platform into their business. So, this was over the last six or seven years, and now you see the wave coming. There are only a handful of groups out there that really do AI from the ground up and build everything themselves. We’re one of them and you just really see a great opportunity for the next 20-30 years of just focusing on this.

 

Tee Ganbold  4:08  

So just to step back for one moment, for those who are listening and cannot see, I’m going to visualize what I essentially saw in your technology. Your technology is able to understand when something is moved from a shelf and has been misplaced or has been removed. Is that correct? Is that what your technology does?

 

Kai-Shing Tao  4:32  

I would say that’s one of the technologies that we do. What we offer businesses is a platform versus a point solution. I don’t think customers want to handle say, for example, 20 different vendors that have 20 different solutions. So we’ve been fortunate in the last number of years for a number of industries—like retail, construction, or agriculture—is that we provide them a complete platform on how you can use AI. So, as it relates to what you’re saying, yes. You know, I think being able to capture missing inventory—which most retailers would consider to be their highest opportunity costs—is a great concern and something that there haven’t been too many solutions in the market to address, and ours does that. And the important part is that we’re live, we’ve been tested, we’ve been doing it for a couple years now. But it also relates to other parts like how do you understand each shopper’s behavior when they’re coming in? And where are they spending their time? But not just where they’re spending the time, but where are they specifically looking on the shelf, to get much more granular in providing that data to not just the store, but also to the CPG companies. So that’s something that we’re excited to introduce to the industry. And all the way to handling the security issues where we now send them moving into the post-COVID area where everyone’s trying to still grow their revenue, but then they really need to decrease their costs, and certainly I think AI is the only way to do that. So retail is a great industry to apply all this.

 

Tee Ganbold  6:26  

So during COVID, I mean, security is a really interesting cost center for, you know, items being stolen. Especially, you know, that’s something a security guard might do, but actually, it’s much more efficient for your platform to help with. But has that been a massive sector of growth during COVID? And will it grow more as more people are financially less stable? Is that what you’re predicting?

 

Kai-Shing Tao  6:59  

That hasn’t been as much of an issue about people stealing, but certainly what is more of an issue is if a product is sold out and no one replaces it. And certainly, you do have the stealing part of it, unfortunately. But as far as for our customers, it hasn’t really raised the flag per se, but our algorithms are able to detect suspicious behavior. The other part for security is just that there are so many new operating protocols to be open, whether that’s the people counting as it relates to capacity, as it relates to PP detection. And there are very significant fines being imposed on retailers if people are caught with that. So our AI is able to spot that and also record it just to make sure that there’s no controversy in who’s doing what or not doing what. And so we provide that, and a lot of our retailers are like, “Well, what’s the cost of not using [our] algorithms?” An inexpensive fine or a lawsuit where things aren’t properly documented, you know, creates an issue.

 

Tee Ganbold  8:13  

Right. And so, again, just to go back, who’s actually using you? We’ve really gone into retail but, as you mentioned, there are various other sectors that might be able to use your solution.

 

Kai-Shing Tao  8:30  

I would say, certainly our biggest customer is China Mobile. So you know, China Mobile is now the largest telecom mobile company in the world where we won the contract to transform their 18,000 stores into smart stores. So we’ve been going under that transformation. It took us almost two years of testing to earn the trust that, one, our technology works, and then, number two, it can actually scale in our deployment. Since we won that, we’ve now expanded into other verticals that have big retail presences. Say, for example, within the Bank of China, with their retail branches, to the China Construction Bank. So, in any situation where there’s a large touching of customers, touchpoints with customers, our AI’s able to come in, understand who the customer is, what they want, what their preferences are and you’re usually able to lower the human costs and cut it down by 30-40% at the minimum. So, we’re now in a bunch of these different industries on retail, as it relates even to inventory control management. Like, we work with a very large coffee chain where they need help counting the coffee cups and caps to the capsules, so our computer vision is able to help them as they— With a lot of these coffee chains, they manage the milk that they purchase, they manage that very closely in terms of the expiration date of when they purchased, so to be able to help control that process better. They take a lot of time and human effort and there’s a lot of human error associated with that. So, AI helps with that project. So retail, construction, and agriculture, we have specific platforms to address each.

 

Tee Ganbold  10:37  

Awesome. And during COVID, which segment or customer may seem to have been pulling you even more? Or where did you get the most calls from?

 

Kai-Shing Tao  10:47  

Definitely the travel industry, the hospitality sector because they need to know how to show, how to make their customers that come feel safe. So there’s a couple points, one, you need to make sure that your employees are safe. You know, tested, safe, gone through all the right procedures. And then obviously the other guests that are there. How are they feeling comfortable standing next to each other or near each other? So definitely the travel part because that trust factor has to be brought back. So our business was certainly in Asia, specifically in China, where everyone knows the adoption of AI is around everywhere. What COVID did for us was be the big spark in other parts of the world where they first used our thermal temperature detection devices. And as they realized that, “Oh, there’s so much more to just taking a person’s temperature quickly,” you know, how you tie that in with the employee authorization to whether they’ve gone through proper testing. There are so many different factors to that and our AI is able to tie all that stuff together.

 

Tee Ganbold  12:14  

It’s incredible. You’ve identified all the needs in China, the most sophisticated AI economy in the world. And now you’re essentially— And you’ve tried and tested, you’ve actually helped some of the largest companies. Now you’re bringing it to the west, which, I mean, the US is trying to constantly keep up with Chinese companies. For a potential customer—American, let’s say, large, it could be a retail chain—probably listening to this and thinking, “How do I know that Remark is the company I should be working with?” Besides all the incredible cases you have, how do I know?

 

Kai-Shing Tao  12:21  

I think that’s a great point because, outside of China, people think businesses love the idea of AI but have been hesitant to take the next step. For us, I think the most important and really, the only way to see just how good you really are, is to go through a proper proof of concept and everything. Everyone can make kind of neat-looking videos, everyone can kind of make the same claims, but everyone can test well in a lab environment and the perfect condition. But when you’re doing it in a real-life situation, and when you’re really testing it with a lot of people, that really separates who’s good or not. Our platform, because we’re early in developing this, everything we kind of offered to the customers is something that we own and built from the ground up. I think a lot of the competitors out there, because of the rush to the market, they have bought a bunch of kind of off-the-shelf products and kind of grouped them together and called it their AI platform. But that’s very hard to customize when you don’t control what you built or what you’re marketing. So I think that’s one reason I would just say, for example, with our retail business, why it took so long to go from technology to commercialization is that they did test us very extensively. So when they wanted to have an idea of who was walking through the store, they would have 4-5,000 people come in through all the different doors where they’re not looking at the camera, where they have different weather conditions that might affect the camera imaging (from fog to rain to sun), they might be wearing a hat, masks, sunglasses, and we needed to be able to make sure that all of our technology worked with all the different people coming in under imperfect conditions. And then you got to test it, and then in an even more real situation, and then we are where we are. So when we expanded into the US, a lot of our customers had the comfort that we— that they weren’t the lab rat. They weren’t the guinea pig. This is something that we have been working on for several years and it was proven.

 

Tee Ganbold  15:15  

And what is the makeup of your team? Is the majority of your team based in North America, are they based in China? Are they in Romania?

 

Kai-Shing Tao  15:27  

Our R&D team is based in the UK, our corporate team is based in the US, and our sales and support tech team is based in China.

 

Tee Ganbold  15:41  

Got it. And are you going to grow your sales team in the US?

 

Kai-Shing Tao  15:46  

Yeah, for sure. That’s a big plan for ’21 and ’22. We got, I think, in ’20 we were able to secure the customers in a number of different verticals and begin to build our brand name in the US as an AI technology that actually works. And so we’ve had a lot of just inbound requests and while it’s hot we certainly want to go as fast as possible. So we really see the US— We definitely even see Europe as a great opportunity. Europe, in general, I always thought that would be slower to adopt on the AI side. But I’d say in the last six months, we’ve received tremendous demand from Europe, even parts down in the Middle East and in South America where they haven’t been exposed to so many different AI platforms or technologies, we see us not doing there. So definitely our expansion plans for ’21 are outside of Asia.

 

Tee Ganbold  16:49  

Hopefully, you’ll be traveling— Well, hopefully, we all be traveling soon to make that happen as well. But going back to some of— A lot of folks who come on here are also somewhat technical or, for those who are listening, trying to learn about the commercial side of some of the cutting-edge companies out there. But for the technical folks, can you talk a little bit about whether your algorithm is training itself? Or is it self-learning? Where are you at in the technology?

 

Kai-Shing Tao  17:26  

Yeah, so I’d say we’re a combination of both. Certainly, the next wave of AI is self-learning and self-training. For us to get to this point, we certainly did a lot of that training ourselves. Unfortunately, it’s just not scalable, so we knew several years ago that, in a way, that party was going to end and we needed to be able to teach the algorithms to train themselves. So that’s something that we have been working on, certainly working on now, and over the next year or two that will be a much bigger part of what we offer.

 

Tee Ganbold  18:06  

Exciting. And just to sort of round this up—and this is something that I continue to do—is ask, what are you excited about in the next 10 years where the supply chain is becoming increasingly automated, every component of it from the retailers to eCommerce. We’re buying so much, but it’s all being integrated. What is most exciting to you and your company with the changing trends right now?

 

Kai-Shing Tao  18:42  

Yeah, it’s just so open right now. If you look at the big trends over the years—whether it’s PC to mobile or PC to the internet, to the mobile, to the cloud—a lot of those, the huge kind of growth curve is not as much there anymore, even though it’s such a big pie and still continues to grow. For AI, we’re so in the first inning right now that just the excitement of companies now actually putting the budget together, having the commitment to incorporate AI into their businesses is exciting to me because we’ve been talking about this for five years and it’s always the same result, which was, “We just don’t have the budget yet. We’re gonna push it off to next year.” There’s always something that takes precedence.

 

Tee Ganbold  19:46  

Pandemic!

 

Kai-Shing Tao  19:47  

Yeah, boy I like this thing! But then money came and surprisingly, and certainly unfortunately, that it took something like this pandemic to give the spark to a lot of this. But we certainly think that our platforms are able to help a lot of businesses and then help the people so let’s do that as well.

 

Tee Ganbold  20:12  

My God, the one book I would recommend to anyone is The Technology Trap, which is basically— Those who don’t want change represent the status quo, and those who want change always use the argument of lamplighters. You know, lamplighters used to go around putting on lamp light before electricity came about. And then after electricity came about, suddenly you can turn on 15,000 lamps in one switch. Lamplighters used to cause so much damage, their job was so dangerous, so repetitive. I mean, we’re all sitting there thinking that was a thing? And hopefully, with the automation and what you’re doing, we’ll look at some of the problems and inefficiencies you’re solving for and think, “My God, we’re creating much more value in what we’re doing now than turning on something over and over again.” So, I just want to thank you for your time and I hope that you get a bit of sunshine and make the most of your day. I’m really grateful, Shing, for you coming on Automating The Chain. This has been a real pleasure.

 

Kai-Shing Tao  21:30  

Thanks for having me. I appreciate it.

 

Tee Ganbold  21:33  

Thanks so much for listening! If you’ve enjoyed this episode, please leave us a review and let us know what you liked. To follow along with future episodes, be sure to subscribe to the podcast platform of your choice, or head over to AutomatingTheChain.Com for the latest updates. Until next time!