in

AI and Digital Grunt Work – Should You Trust Predictions? | Interview with Rett Crocker (Part 2)

How do businesses use AI to become more efficient? How can we better utilize A.I. today? How does AI differ for business vs consumer use cases? How should a business leader construct their team with AI in mind?

This is part 2 of our interview series we had with expert Rett Crocker, CEO/CTO of Udu. See part 1 here: https://youtu.be/qwu47cjVW70

Learn more about Rett and Udu at https://www.udu.co/

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Alexander Ferguson 0:01
Welcome to UpTech Report series on AI. I’m Alexander Ferguson. This is the second part of my conversation with Rett Crocker, CEO and CTO of Udu in Raleigh, North Carolina. Rett has designed and developed over 100 games for mobile devices, personal computers and video game consoles. He’s also invented multiple programming languages, game engines and multi user content. And he’s created innovative software technologies in fields ranging from speech synthesis, to adver. Gaming to collaborative education. I asked Rhett what examples he seen of businesses using AI to become more efficient, how AI differs in business use and consumer use, and how business leaders should construct their team with AI in mind.

Unknown Speaker 0:47
So I’ve seen a few examples, not nearly as many as I’d like to see. So two examples that I will give both relatively short one is one of our customers, one of our oldest customers, they track apartment rentals, availability pricing. They, they the way they used to do that, well, two thirds of the problem is solved. Because two thirds of the market is these big giant apartment buildings that are owned by big companies that have tech teams that build an API that they can just hook into the the last third is the hard part. And that the way that last third works is they have humans and cube farms going to websites and saying there’s an apartment available on the first floor, it’s $2,700. And it was $2,600. Last six months ago, square footage is the same it looks like they say it’s got two and a half baths now instead of you know, one bath, so like what’s going on, I don’t know, they have humans doing that problem. They now use you to to do that work most of the time. So for about half of that 1/3. I guess when sex for about one six of their overall catalog, you do just goes and collects all that data for him. There’s no humans involved at all. For the other six, there’s some small percentage that you can’t help with at all. But most of the other examples. This customer has those same cube farm people. But instead of going to those web pages and entering the data themselves, and only getting through like basically 20 apartment buildings a day, which is what their rate was. They instead go to that apartment building website and they click a button that they call the YouTube button. And it basically sucks up that website and sends it up to you do you do processes, it does all sorts of crazy stuff with the text, and then shoves that data into their CRM. So all that humans therefore is clicking that button. And figuring out what the right pages in the first place and doing human things that are actually like more difficult. They, I think, I would argue their quality of life is better, although not much better, in my opinion, but has a different issue. The second example that I would give is there’s a movement I’ve definitely seen in in the private equity space just because of the fact that we’re currently serving that market of people doing what’s called deal origination or deal sourcing. Using more automated methods, some are using people like us, there are some that are trying to use different AI techniques. I haven’t seen many great examples of them being successful with it except for with our stuff and a couple of the other competitors that are actually mostly using humans. But the thing that they’re all trying to do is basically to make it so that their humans don’t have to do that particular groundwork again. Because the way that used to work is that they would buy a list of all the like they’re going to buy, they want to buy a number of dentists, dentist office, dental clinics, and they’re centered around the Northeast. And the way they would do that is they go to Dun and Bradstreet or you know, whatever, somebody and they’d say, give me a list of all the dentists in the Northeast or they look into phone book or whatever, right? And then they’d have MBA interns, people that actually make real live money, like sometimes six figures, going to each one of their websites and saying, how many dentists that I have? How who’s the senior guy the owner? Oh, how old are they? Are they over? 50 Are they may be getting ready to retire. Maybe they’d be interested in selling all those things. That’s what Humans would do every day day in and day out and have done for years and years. But they don’t do that now. Or at least every one of those people is trying the smart ones anyway are trying to figure out how to not do that, because it’s such a colossal time sink,

Alexander Ferguson 5:18
how can business leaders better utilize AI? And what are the potential issues those business leaders may face.

Unknown Speaker 5:26
So I would use it in cases where you need to predict the outcome of something, that’s the best case scenario use for it. That’s what it’s good for. That, or automation are the two big uses. So an example of the automation I gave earlier, the, you know, humans sitting at the desk, you know, clicking the YouTube button and uploading the website, and then you do does all the processing on it. So the human doesn’t have to do data entry. That’s a great example of automation. The prediction side is, you know, if you’ve got a business case, where you’re trying to predict how much your customer is going to spend next year, or whether or not customer x is going to like something or not, those are great uses for AI, because you’re basically saying, I’ve got a bunch of data that shows how people have reacted in the past, and are going to use that to try and predict the future. The biggest difficulty relating to doing anything sort of AI predictive is getting good data. That is that is the hardest thing. And then also not believing the hype around it. Because the way the math works, you can end up finding sort of false results pretty easily, that are that seem great, but aren’t real. So it’s very important to build all your systems, all your models in such a way that you can test them and verify that they actually work.

Alexander Ferguson 6:58
How does the threshold for success change? When AI is being used for business versus consumers?

Unknown Speaker 7:06
And business, you only have to get 80. But in consumer, you got to get closer to 99? Which is why people are you know, Desirey because it’s like, oh yeah, it didn’t it didn’t understand me that one time. Whereas with business you can be there, it’s a little bit more pragmatic. You can be like, well, it didn’t work 100% of the time. But you know, eight out of 10 times, I didn’t have to do anything. Microsoft Office Microsoft Word back in the day had the paperclip? Yeah, I see you’re writing a resume, you want some help with that? No, Clippy go away. The the the reality is that feeding that sort of information in is quite. And having a try and predict and help you can be quite useful. But I think that one is sort of a consumer level, like bar that you have to reach where it’s like, it’s got to be right 99% of the time, or you’re going to be really annoyed with it. And you’re gonna turn it off. If you can use AI, and the principles of statistical analysis and predictive analytics and all that to get you 80% of the way to solving your problem, then you’ve got something very valuable. Even if you have to use humans for that last 20%. That’s still dramatically better than having your humans doing that. Do 100%

Alexander Ferguson 8:24
How should business leaders construct their team with AI in mind?

Unknown Speaker 8:28
First off, I wouldn’t build anything from scratch anymore. There’s just almost no point in that. Because of all the existing tools that are out there that are open source. Thank God for open source, right? It is such a useful thing in problems like this. It’s the same value that sort of the open academic community brings, right people trading knowledge and stuff like that. Ultimately, what you want are good programmers. That’s really what it comes down to. Preferably good programmers that understand math, because math has kind of a key component to it. But there’s room for both the ones that are less mathy. My team, I literally, I literally put ads on Craigslist and interviewed people, for the most part. The ones that have like my oldest employee was was the very first Craigslist ad when I was like, let’s see if it works. Let’s hire a few people and see how this plays out. And he came in and he had zero programming experience. He was a mechanical engineer, actually. And he was like, this is interesting. I’m interested in startups. I give him the pitch on you do what it what it was and what we’re trying to build. And he was like, Yeah, I’d like to do that. And I was like, huh, how’s your programming? He’s like, it’s not great. And I was like, okay, but he was smart, and he could problem solve, and that’s part of the key thing that I one of the reasons I bring him up is the right people to solve these problems in the future are good problem solvers, because there’s all these other tech analogies that you can integrate. And you can do a lot of different things. And you can solve a lot of different problems, but the real to build good models, predictive models, you need to build good data. And, you know, the sort of dirty secret of data science and AI, which is part of that is that folks that call themselves data scientists, which is really just programmer slash statistician. They, they spend most of their time cleaning up and finding data, which is gross, terrible, no human should be doing that is one of the reasons we, you know, point you do at that problem all the time. It’s like, you can do that for you, you don’t have to, but if I’m, if I’m some, you know, random CXO, in, you know, corporate America and I need to apply predictive analytics or artificial intelligence or what have you to try and predict my, you know, customer, how much my customers are going to spend in the next year or whatever. You know, 90% of the time, if I was building a team to do that, I would get smart problem solving problems. Maybe a data scientist,

Alexander Ferguson 11:15
back includes the audio version of this episode. To see the original and more visit our UpTech Report YouTube channel. If you know a tech company, we should interview, you can nominate them at UpTech report.com. Or if you just prefer to listen, make sure you subscribe to this series on Apple podcasts, Spotify or your favorite podcasting app.

SUBSCRIBE

YouTube | LinkedIn | Twitter| Podcast

Breaking Down A.I. | Interview with Rett Crocker | Part 1

AI Augments Real People. Sort of… | Interview with Rett Crocker (Part 3)