Editor’s note: In the third of several in-depth interviews with Alexander Ferguson, CEO and founder of YourLocalStudio.com, talks with Richard Boyd, founder of Carborro-based Artificial Intelligence firm Tanjo about AI – what is it, what can it do, what it means for the world in years to come as a disruptive force.

The video is part of a series, UpTech, from YourLocalStudio which is partnering with WRAL TechWire to publish the series, including full transcripts of the interviews.

Welcome to UpTech Report series on AI. I’m Alexander Ferguson. This video is part of our deep-dive interviews, where we share the wealth of knowledge given by one of our panel of experts.


In this episode, we continue our conversation with Richard Boyd, founder of Tanjo in Carrboro. Richard is a successful entrepreneur, author and speaker, so we wanted to know how has his thinking on technology and AI evolved.

  • How does he use it in his own business? And how could other business leaders use it in theirs?

Early on, our team is working in computer gaming. It started when I met David Smith here in North Carolina. When I met him, he was beginning, he had just done a game called The Colony, which was the first real-time 3D adventure game that attracted a lot of attention from people like Tom Clancy and guys like James Cameron, who at the time was working on a movie called The Abyss down in South Carolina.

So just this idea of taking technology and applying it to problems to solve them like what we saw, we helped James Cameron solve some visualization problems around the movie The Abyss, early on, and that was a fascinating process. But I guess today, so I guess it was a natural evolution, right?

Like applying technologies to problems and we ended up getting really interested in Artificial Intelligence as a way to build deeper meaning into the virtual worlds we were building and like I said, with computer environments, building more convincing characters that you can believe in more convincing environments, and it sort of just evolved from there.

I mentioned 2009 as the time when we got religion, so to speak, on machine learning. And that’s when David and I went out to Microsoft Research Labs where Alex Kitman was working on the Microsoft Kinect. If you remember that, it was called Natal at the time. They were just trying to teach, so they were kind of this, I understand that they don’t sell it anymore, but there was a piece of hardware you could attach to your Microsoft Xbox that would watch you in your living room, and you could use your body as the controller. That was the central idea.

But in order for that to work well, the sensors had to be amazing. So they called Lockheed Martin, and Lockheed bought my last company, so David and I were there, we went out to Microsoft Research Labs, and it would happen to be during the Game Developers’ Conference, we were out there anyway.

Walked in and saw Alex and there’s a guy named Jaron Lanier there who I’ve known for a long time. He’s a guy who came up with the term virtual reality. He’s this kind of dreadlock guy. You see some pictures of me online with, and initially we were looking at how can we help you with the sensors, should it time of flight, should it be structured light, whatever. But we found out pretty quickly they had that nailed. And what they’d built was like the optimal solution to that problem in the form factor that they had to fit it in.

But the other thing that they showed us was, oh yeah we’re trying to teach this system what a living room is. And that is a difficult computational problem. And so, again, there were two approaches. The approach at that time still coulda been, let me just program in and tell them what a chair is, what a table is, what a plant is, a whatever. Or the other way, which they thankfully used, was machine learning. Which is, let me just have examples of living rooms from all over the world, Asian, European, South American, U.S., rural versus urban, whatever.

Everything you might encounter in that, and give them all the, give the system all those examples. Millions of millions of examples. And whatever they did had to fit within about less than 100 megabytes of space. The whole brain for the system. And they were able to achieve that. So that just blew us away, and that changed our thinking completely.

  • How did that revelation change the course and direction of your business?

– Secretary of Education at the time, Arne Duncan, and his Deputy, Jim Shelton came to Lockheed and said, hey, we in the government have lots and lots of information at the Smithsonian and all over the place, in the Library of Congress. How do we make it available to teachers in an easy way?

Like we’re trying to scan and digitize this stuff in, but how do I make it discoverable by tagging it? Right now we’ve got armies of human beings in there trying to put tags on stuff. And I usually use a picture of that last scene in the Raiders of the Lost Ark, where you’ve got a clerk with this crate, and it says Ark Thingy on it, and it’s the Ark of the Covenant, that can destroy or save the planet, you know and it’s inside this box. He’s putting in this massive warehouse with a tag that says Ark.

It’s like, that’s undiscoverable. It’s a potent thing that’s valuable, but undiscoverable. And most of the information we have, is what we call dark data.

Inside Artificial Intelligence’s disruptive power: From marketing to resurrection of characters, singularity, more

Right, it’s whirled away somewhere, inaccessible and undiscoverable because it’s not digitized or not tagged. Even if it’s digitized, it’s not tagged. Well, what’s amazing with, what we did for the, what’s called the learning registry for the Department of Education, was built a system that can go and look at that stuff, you could read the Declaration of Independence, or the Magna Carta or any other document, or look at images of things, and if it had something similar to it in it’s massive multi-dimensional lookup table, it would go ahead and tag it, right?

And if it didn’t recognize the thing, then it would say I need a human expert, and it would call for help, right?

Phone a friend. In this case, it’s a human, to come in and say, oh that’s actually an ancient Cluniac drinking vessel from 100 B.C., and go ahead and tag it. But of course, once it’s been tagged once, the great thing about machines is they never forget.

  • How can businesses and organization implement practical A.I. applications?

So whether you’re practicing law, or you’re practicing architecture or whatever, what you want now is a machine learning brain like I’ve just described, that goes through everything that you have.

All the assets that you have, and reads everything. Every document created, every, ideally, to be honest, every email written by all of your people, and it understands like what do people know? What is our organizational knowledge?

And it maps it all, and by the way locates where everything is. Which is something that’s incredibly important for digital transformation. And then once it’s mapped, now you can track things like how does new information enter organization? Who’s championing it? Who’s challenging it? How do decisions get made? Why did we choose this vendor over that vendor? And why did we choose this strategy over that strategy? And we’d be able to tell you forensically, what decisions are made and why, and maybe help you make better decisions in the future. What made Red Hat really successful, here again here in the triangle, was this idea that implementing Unix based servers within your organization is a complex activity. It’s also very intimate activity, because it’s where all your people are connecting, right?

Deep dive: Artificial Intelligence can take over ‘things humans should arguably not be doing’

So do you wanna just outsource that to someone else, or do you wanna buy turnkey on premise tested solution that works extremely well that you can shepherd and manage going forward? And that was that decision that led to how much should they just get bought for 34 billion or whatever it was, right? From free software. So I think that same principle applies here. Is that again whether you’re a government organization or you’re a company, you’re an architecture firm whatever, get your own machine learning system inside your firewall under your control, and make sure you know where your data is and where it’s going.

If you wanna get people fluent and comfortable with this idea that hey I want this asset here, I want this machine learning companion that’s gonna help me do my job better, but also it becomes an enduring sort of map of how decisions and how work is done within the organization, so we’re doing that for all of North Carolina Community Colleges in North Carolina, so there’s 58 of them, right? And when this is fully implemented, one of the things that I, ’cause I’m on the board of trustees of White Tech, right?

So I understand how turnover happens, and the sort of complexity of some of these organizations. At White Tech we have like 70,000 students a year. It’s a 250 million dollar enterprise, that’s underway that does a lot of good in the community. But we just, our president just, like presidents do, they retire. So now we gotta put a new person in place. And there’s lots of other turnover that happens at a various levels throughout the organization. What if when that new person steps in, they could see right away, like they have a little companion A.I., that assistant we talked about earlier, that says, well it looks like you’re approaching your first board meeting.

Well the last person, here’s when they approached this kind of problem, here’s the resources they went to, here’s the people they went to, and here’s how they did that job. It’s also really good for the organization to have that have a map of that intelligence within the organization when people leave your company, and you’ve invested a lot of money in those people. And time, you’d like to have some model of that that stays behind after they leave.

  • What kinds of success have you found with using A.I.?

Our entire solution set is around something you can implement in less than six months that will have a 10X return on investment. And that’s our kind of guiding algorithm for everything that we do. So that means that we’re looking at the low hanging fruit, things like accounting, so we work with a local accounting firm, found out that there’s some new rules around revenue recognition and lease recognition that were coming out.

We looked at that and said that’s perfect. Because all I need to do is get a bunch of sales contracts, feed it to a system and don’t let it see what Kanye West or anybody else is doing, nothing else on the internet. Just focus on this very tightly bound realm of what kind of language will you encounter in a sales agreement?

Taking a deep, deep dive into Artificial Intelligence: A new series to watch, read

Whether you’re selling software, hardware, services, consumer goods, whatever it happens to be. And become familiar with all of the terms and everything, and not just like 10,000. Not 100,000, but millions of contracts.

Let it read all that, create its own sort of inferred understanding of how to process that, and then just do the basic job of bin sorting, like yeah this is a really standard contract. This one, these have a few non-standard elements that a human being needs to look at. These are on fire. Like this is all non-standard.

Whoever’s doing this is probably trying to cheat you, and this needs a lot of human attention, or you probably should not do business with that person. Whatever the rules are, right? And so applying it to things that are easy to digest, and get that 10X return in a short amount of time, that’s where we find our success.

This was just a taste. Stay tuned as we share the full deep-dive interviews we had with each one of our panel of experts, and our upcoming episodes, focused on specific topics that will transform the way you think about Artificial Intelligence. All this on UpTech Report’s new series on A.I.

Artificial Intelligence and you: Introduction to a new video series ‘UpTech’