Future of Receivables Management | Rob Grafrath, Grafrath Consulting | Ep. 245

Adam Parks (00:08)
Hello everybody, Adam Parks here with another episode of Receivables Podcast. Today I'm here with an extra interesting guest joining us to talk about the future of debt collection. And part of the reason that this guest is so interesting to talk about the future is he is actually an accomplished science fiction writer talking about the future and being able to formulate those types of imaginative ideas.

Rob Grafrath (00:27)
So I

Adam Parks (00:36)
and apply them to a world in which he has worked for, I wanna say close to two decades in the debt collection space. So if you have not met Rob Grafrath yet, he's now out there working with a variety of clients and I think it gives us a really interesting perspective for today's discussion. So Rob, thank you so much for joining us, coming on and sharing your insights today.

Rob Grafrath (00:43)
Yes.

I'm not sure if we're going to

Thank

Thanks for having me out. I'm excited. So excited to be here.

Adam Parks (01:01)
Absolutely. So for anyone who has not been,

well, we're excited to have you. For anyone who has not been as lucky as me to get to know you, could you tell everyone a little about yourself and how you got to the seat that you're in today?

Rob Grafrath (01:13)
Yes, sir. So yeah, probably 26 years, I'd say that is the number I started as a debt collector in 1999. In Kansas City grew up in the one and raised in Kansas City did that for just about a year and a half before I hopped over to the IT side of the house ran development ended up becoming an IT manager as well moved down to Texas company up in Sherman. And now I'm in Fort Worth. So from pretty much all of my career. It's been the traditional salary positions in IT positions, but it's also been leadership. And I've had the benefit of having some great mentors over there on the operations side and just being in the room and talking to everyone about their business and how it works. And actually going to shows and paying attention to the industry as opposed to just being the IT guy that just kind of wallflower, give me some. Give me some requirements and let me go. You know, I have to understand the business reason behind things. So then, yeah, in January, I broke out to do consulting work. had a taste for it, did some subcontracting a number of years ago, and decided this was going to be the next full-time thing for me.

Adam Parks (02:17)
Well, that's fantastic. Your experience and ability to understand the business requirement behind the technical need is important as we start moving into the future of what debt collection is going to be and how do you operationalize some of these really great technologies and ideas. And as you and I both attended the ACA Conference and the RMAI Executive Summit and getting a better feel for what's happening in the space right now. I'm curious to get your thoughts on some of the technology and tools that you're starting to see enter the marketplace.

Rob Grafrath (02:53)
That's what we're here for. going to talk about the future. We're going to get a little sci-fi today. ⁓

Adam Parks (02:57)
Well, you know, let's take one step back and let's go through the sci-fi adventure as well. So could you tell everyone a little

about your books and what inspired you to start writing?

Rob Grafrath (03:06)
I sure did.

Yeah, I've always just loved putting words out into the world and it probably was something like 10 years ago. I just buckled down and said, I'm going to write a novel. I wrote my sci-fi novel. I stole liberally from the personalities of my family. got four kids and my wife and even our dogs might have a little pieces to play in that book. Started as one way too big book. I broke it up, turned it into a trilogy. And then did the self-publishing thing to get it out into the world. Didn't want to have to ask a publisher's permission to have my words out there. So it's been very much a learning experience of how to structure that story and actually trim it to make it because producing words is easy. I can throw words out there all day long, but actually Trimming it down, editing it, making it punchy and keep the reader's attention is a whole other story.

Adam Parks (04:02)
laying down the foundation of the ideas and then clearly communicating those ideas really is two different worlds. The writing versus clarifying. So as we talk about the future of the debt collection industry, I think we could make the argument that the future is right now.

Rob Grafrath (04:13)
Yeah.

Adam Parks (04:20)
We're starting to see some of the technology that we imagined in the 80s and 90s are starting to come to fruition. It's starting to feel a little Jetsons out there minus the floating cars. But talk to me a little about, I would like to find one. I feel like that might help with some of the traffic issues in South Florida. But what kind of things are you seeing in the marketplace right now that are making you feel like that's a true statement the future is now?

Rob Grafrath (04:21)
That's right.

Where's my flying car? Still waiting.

Thank you.

Yeah, you can't turn around without AI smacking you in the face, of course. But what's going on today when you see AI doing, I think one of the unexpected things is when AI was fledgling and we were thinking about the future in the past, when we were thinking about the future, you were thinking robotics and flying cars and stuff like that. Service jobs seemed protected it seemed like when early computers came out and people were imagining what what the future of AI might look like you'd have Even technologists people trying to make predictions. They're like, it's never gonna beat a human at chess. That's absurd It's way too complex and sophisticated. It's a human endeavor and then when it did Then then go was next alpha go Everyone was like no one's gonna be a human at go. That's just there's more possible moves than there are atoms in the universe. You can't you can't beat a human to go and then it did. So then looking at AI passing the bar exam, AI diagnosing patients better and reading MRI and x-rays more accurately than human doctors are and so the service industry and these you know programmers right as a programmer I know

Rob Grafrath (05:54)
AI today is my friend. Will it continue to be my friend five and 10 years from now? We shall see. You might argue that even today programming jobs are being lost to AI. You might argue that if you consider that perhaps the typical programmer is more effective and efficient in what they do. So you don't need a staff of 10. Maybe you five, six, seven. So that translates to possibly lost jobs.

I think it's interesting when we start talking about, know, are we going to be able to move away from that? I've had a couple of conversations recently where we were talking about AI being the amplifier, not the replacement, right? We're not going to see artificial intelligence in the near term start taking over complete jobs. But I think what we are going to start seeing is an amplification of the individuals that we have and enabling us to focus

Rob Grafrath (06:21)
We like to think that it's a shift.

Adam Parks (06:45)
our real people on those instances that require empathy or special circumstances or for us to walk a tighter rope on the phone with a consumer for whatever reason it may be. But that seems to be the general gist of how the industry is looking at it right now. It sounds like you're you're pretty much on board with that assessment.

Rob Grafrath (07:06)
Right on. Absolutely. And the technology, it's a tool. It's how you use it that matters, right? Just like any technology. It can be for good or for bad with Doge coming through, right? And assessing all of the government regulations. There was an announcement that Doge, they had their AI deregulation tool. They've let it run rampant to read through and process more regulatory guidance and more documents than any human can within a lifetime, right? And then decide which of these can be trimmed, which of them need to be retained in order for that body in order to keep functioning. And that's, you could argue that's a good positive thing, perhaps, to trim some of the fat, right?

There's maybe others that would argue otherwise. That's one of the things with, again, the good and the bad is like technology, could say from one perspective might be technology in the wrong hands and then from another person's perspective, that's how the technology should be used. ⁓ If you've heard of in Iran, they have some very strict laws around women wearing the hijab.

Rob Grafrath (08:17)
and being in public. they recently read a book where I learned of the AI face recognition technology being deployed in Iran. And if you're a woman driving a car, you get a text without a hijab, you get a text message that you're violating the law and your car is going to be repossessed, your license is going to be suspended, like you're in deep trouble. ⁓ There was a lady in the parking lot putting her groceries away in her

Rob Grafrath (08:42)
a job slips down off of her face and then just a few seconds later, text message comes across. She's, she, they caught her. Like you're in a parking lot. Like how does that even happen? But drone surveillance and AI face recognition technology in those hands, you know, from their perspective, they're, they're enforcing their laws, right? So you could say, well, it's just enforcing the law. It's, it's a matter of, but some, you know, human rights activists would, argue otherwise.

I look at it, I think about it, you know, to circle it back to the debt collection space, I think about compliance and compliance calls. So there was a time where listening to five or 10 % was a reasonable volume of calls to be listening to in terms of your agencies or your agents, depending on how, you know, where you sit in the cycle. ⁓ I think there's some interesting challenges to that. And the application of artificial intelligence to bubble the potentially problematic

Rob Grafrath (09:11)
And then they put in, you know, the district loan package, the debt collection to me, and I looked at the money and said, that's great. So there was a flag retention. So, I think this was a good picture.

Adam Parks (09:38)
calls to the surface so that the live individuals are able to check the right five or 10 % of calls that need to be checked is a little bit of a different dynamic. I see that as being kind of at least in the near term, what the future is going to look like us as an industry for amplifying our capabilities to do things without necessarily turning it into full automation. I like to argue that debt collection companies are only 15 % unique.

Rob Grafrath (09:42)
and make sure you're still in the zone.

you

So, you're true.

Adam Parks (10:08)
And I believe over the next, you as we move from the near term to the, let's call it the medium term, I think one of the things we're going to start seeing here is this.

Rob Grafrath (10:12)
yourselves.

Adam Parks (10:17)
movement to not necessarily full automation, but I think you're going to continue seeing how are we going to keep that human in the loop and how are we going to make good decisions going forward, right, with some sort of data set.

Rob Grafrath (10:19)
And I was concerned about the office. More empowered use of these.

Adam Parks (10:32)
So how do you feel, how do you think that the consumer is going to interact with these newer technologies? Do you think that the use of technology in the debt collection space is going to trail how the consumer preferences change or do you think consumer preferences will start changing based on how we're communicating?

Rob Grafrath (10:50)
It's a give and a take. You have both consumers being empowered with new technologies and our industry and other industries being empowered with same and similar technologies. where consumers, you've heard of the new iOS updates that we'll have some call screening and some call summarization technology available to consumers. you can expect, and this might be getting into that near future discussion.

Rob Grafrath (11:13)
You can expect that, consumers, they'll have their AI bots, they'll have their agents that call on their behalf. And then now we were introduced to a whole new set of problems. How do you authenticate that? How do get that permission to talk to that bot? And then let's say you get that and they'd rather speak to you through that bot and let it negotiate terms rather than having a direct conversation. That's not a far stretch of the imagination.

Adam Parks (11:39)
gets highly probable considering what we're seeing from a technology standpoint right now in terms of communications. How much of you how much of this do you think will be cyclical? And I've got a weird scenario and I'm curious to get your take on it. So my wife and I have a have an age gap and we communicate with our friends very, very differently. My wife sends voicemails like she texts voicemails back and forth with her friends, right? So she like talks to her phone, she texts it out to that person.

Rob Grafrath (11:40)
Good luck.

Adam Parks (12:04)
I, on the other hand, have not listened to a voicemail since 1996. So do you think that, like, is that an application of technology being cyclical? It's coming back again. Like, they're starting to use the, they're communicating an old school technology in a new way. Do we see more of that in the future?

Rob Grafrath (12:09)
So. Thank

Yeah.

And you can also, I agree, right? That it kind of comes and goes in cycles and every generation has their different implementation and use of the technology. We might not have expected when texting started becoming a thing that all the kids were like, it was back when you had to touch the pad, you had to hit the number two, you know, three times and then hit the number three one time in order to come up with a text message if anyone remembers that. But I see, yeah, that sort of generation by generation usage of technology and the differences in unexpected ways the different generations want to engage is going to drive a lot of what our industry does.

Adam Parks (13:03)
Well, as a debt collection industry, and as we start talking about money and the collection of money, money seems to be changing with technology as well as we start moving down the crypto path.

Rob Grafrath (13:04)
This is a one.

Money itself has changes with technology. Like what is money? What defines money becomes a little bit more fungible with the technologies. And there we are in that near future. Blockchain, right? That's not new. We know that Bitcoin's out there. We know that blockchain is a thing. But I would argue that the average consumer really doesn't use it much, doesn't care much. I don't think that the typical

Rob Grafrath (13:42)
walking down the street person has a bunch of Bitcoin and that's how they're they're engaging in any transactions that they that they undergo. I do see that changing in the future. I think that I think there will be a killer app. I can't can't say exactly what that'll be, but something going to unlock the common use of blockchain and cryptocurrencies for the for the average user.

So it's interesting that you say that I've got some stats for you I think you might like. And so Brazil has a platform called PICS. India has another very similar platform. And that's where most of the transactions are flowing through, meaning it's attached to your bank account, the coconut, the guy selling coconut water on the side of the road, it's got a QR code on his cart. You scan the QR code and make the payment and he gives me the coconut water.

Rob Grafrath (14:07)
Thanks.

Thank

Adam Parks (14:30)
The volume and dollar value of the transactions executed in the country of Brazil will exceed on PICS in 2025 what they did on credit cards in 2024.

Rob Grafrath (14:44)
So yeah, the international communities, it's surprising, right?

Adam Parks (14:45)
we're just seeing the flip.

Because of the the central bank digital currency in the United States, I think will continue to get pushed back against and I know that there's been some, some new movement in the crypto market in the US and being able to leverage stable coins and things of that nature, but the technology outside of the United States right now is moving at a rapid pace. Like you talked about the facial recognition technology that's being used in Iran, in China, and the US airport. So those things, think, all start to, when you start collecting all that information into one place, one, I think it creates risk. But it's interesting to see that adoption speed outside of the United States for this type of technology. I don't know anybody in Brazil who walks around with currency in their pockets.

Rob Grafrath (15:34)
Yeah, kind of like the social media messaging, right? Like the idea of as a debt collector, we want to have more ways to contact people. You have texting and phone and not a whole lot, you know, voicemail drop. But as I've heard in other countries using WhatsApp and Facebook messaging to reach out to consumers to do debt collections is not anathema quite like it would be.

Rob Grafrath (15:57)
in the United States. know that people have done it. I've worked for a company that did some Facebook outreach, but it was very timid. It was very like frowned upon.

Adam Parks (16:07)
I think it's challenging because prior to Reg F we didn't have clear understanding of the rules in the debt collection industry because it's 85 % regulated. I think really does move a little slower when it comes to technology and we have to keep our eyes on what's happening in the lender space. Fintech lending was happening a long time before Fintech in the debt collection space. So I...

Rob Grafrath (16:23)
So

Adam Parks (16:32)
Part of that I think is the regulatory and risk environment. Part of that might be the margins are quite different in debt collection versus lending. And so there's some interesting dichotomy there. But what do you think happens with debt settlement? What about some of these other, let's call it adjacent features of the economy?

Rob Grafrath (16:50)
Yeah, I'm putting that in my mid future category that I believe that the debt settlement industry and the debt collection industry, I can see a merge coming in the future. I think that it could end up being kind of two sides of a coin that you have starts as loose partnerships. And today we have vendors that you can scrape and you can

Adam Parks (17:02)
Interesting.

Rob Grafrath (17:13)
reach in and say, hey, I've got that new debt settlement. Let's proactively work with these debt settlement companies. think that that will merge into partnerships. And then before you know it, debt settlement and AI-driven financial management tools, that'll be like a package deal. And if things don't go well for that consumer, then on the back end, here comes the debt collection.

Adam Parks (17:36)
Fair, I think there's some interesting opportunities and we've already seen some organizations that start to blur the line and cross between debt collection and I'm gonna call it debt relief, because it's not just debt settlement sometimes it's credit counseling or some of the other services that are available to consumers through that environment. Now, Some of the other things that we're starting to see in the market, like I'm a big Apple fan. I started playing around with the Vision Pro because I wanted to understand what this augmented reality would start to look like. And I realized that we're not anywhere near any kind of a commercial application at this point in time, but it's always interesting for me to stay on the forefront of those. But then we start thinking about,

Rob Grafrath (18:03)
Peace.

Adam Parks (18:16)
Metabuy's Oculus, we're dealing with some of these virtual realities and I'll say augmented realities. How do you think or do you think that will start to play some sort of an impact in the world of debt collection in the future?

Rob Grafrath (18:30)
Absolutely. That's a future marketplace. Talking earlier about the different contact venues, you can text and you can email and you can send physical mail. And it's definitely today, it almost feels like a joke if you say, yeah, I'm on the metaverse. Okay, me a break, really. What does that even mean? No one's there in the metaverse doing anything. So I don't put it in my five year bucket. I'm thinking more closer to 10 or plus years in the future. I do believe that a like once the the the VR and the AR technology isn't bulky cumbersome and expensive and and you can just slip on a nice pair of regular old sunglasses and then go to town and you don't have this heavy thing on you. You talked about the counterweight you have to put on your on your how does that work again?

Adam Parks (19:15)
Literally, is it similar to night vision? If you're have a weight on the front of your head, you put a weight on the back so that it balances out and you're not stressing the back of your neck to hold something in place for an extended period of time. But I will tell you 360 monitors while traveling the world is as efficient as I could hope to be from a hotel room.

Rob Grafrath (19:21)
and have a lead on the front panel. Nice. Awesome. there. Fantastic.

Yeah. that's a dream, right? As a programmer, I just imagine I'm just kicked back and I've got, I've got every such as the code here and I got my emails here and I, I'm not confined to these, these two monitors or four, some, you know, some people have more than my sit stand desk doesn't, doesn't do well with more than just the two. but yeah, so

Rob Grafrath (19:54)
You put on your meta glasses or whatever, and then you go into what eventually will become a more robust marketplace. And it's not just this one little meta and their metaverse as a VR place where you can go, you meet, you have a presence. That'll be a way to find contact and interface with consumers.

Adam Parks (20:14)
So one of the things that I've read about and this may be a little bit nerdy, but I'm gonna go down this path anyway. It really is. So I started thinking about Grand Theft Auto as a video game platform, which was, if I'm not mistaken, Grand Theft Auto V was the most popular video game ever created in the history of the world. And I'm sure there's some competition with some of the others, but that's at least my understanding of it. And as they started talking about the new version, which I...

Rob Grafrath (20:20)
That's what we're here for.

Adam Parks (20:42)
think has been 10 years in the making. Some of the rumors were around like, they're gonna create a cryptocurrency that is driven within these gaming systems or within these gaming environments or metaverses. So within those, let's call them self-contained worlds. Do you think that there's ever a place for debt collection within that? Like, do you think that there's going to be lending and buying and, I think there'll be some buying and selling and wherever buying and selling exists, there tends to be lending.

Rob Grafrath (20:54)
and things that we can about. If you feel like that's a word, please. Thanks. It's like

Adam Parks (21:27)
which you can collateralize against it.

Rob Grafrath (21:31)
Wherever debt can be incurred, you will find us in that bright, future.

Adam Parks (21:36)
wherever transactions happen. Wherever two people are going to transact, wherever something is gonna be bought and sold for the most part, even if you only wanna look at it from the perspective of the credit cards in the real verse, anytime that there's buying and selling, there's some sort of debt or service even being provided. Your power company, your water company, your trash company, all of those are services that are provided.

Thank

Adam Parks (22:02)
in advance of and ultimately create some sort of debt. I'm wondering how and if that will have any kind of an impact into the debt collection world of the future.

Rob Grafrath (22:05)
Thank You also think about anywhere that a consumer exists, right? Anywhere where they can be reached is an opportunity for us. as new wherever they are, like if everyone's going to the metaverse, you're going to find us in the metaverse, right? If everyone's texting, you know, as of 10, 15 years ago, everyone started texting. And of course, we didn't have

Rob Grafrath (22:38)
Reg F guidance right out of the gate, but once we we got that that really unlocked we wanted to be texting I think a lot of there were some early adopters that were doing it But it opened the gates and now texting is is just table stakes. Of course, you're gonna be texting But you go back before reg F and it be like I don't know if we're gonna be doing that ⁓ but even if you imagine like hyper targeted ads or something like that like What if this is just a crazy idea? So so stop me if I'm if I'm way off the rails here Bring it.

Rob Grafrath (23:05)
But with as much as granular as you can get on this particular user is Adam Parks, you know, is Rob. That means that I could pay for an ad to show up when Adam goes to search and he looking for something ability unrelated. Let's say he decides he's going to buy a new sports car. But I know that he's in debt from that car that we repot last week. And then next thing you know on the ad the first ad that pops up as ABC collection we'd like to talk to you Adam you it wouldn't be yeah it wouldn't be you have to worry about a third party disclosure you can't can't have your ad say this is an attempt to collect a debt can you

Adam Parks (23:36)
Are you listening to my pitches at ACA?

So, it's weird, you

can't have an attempt to collect a debt, but you can provide education. I can talk about the cost of bad credit, the value of good credit. We can talk about that this organization employs thousands of people in their local community. like, I asked this question of quite a few people at the last few conferences was how much dumb shit have you bought on Facebook or Instagram? And why aren't we leveraging the same kinds of technology? Because there are...

Rob Grafrath (24:00)
to legitimize yourself.

in front of

their faces.

Adam Parks (24:12)
debt buyers and agencies that have been doing that since like 2018. I actually did a podcast with Kristen Leffler from Resurgent talking a little bit about some of the technology that they've deployed and how they've kind of layered on some of the e-commerce technology on top of the debt collection process. Because when you look at those two different funnels, they're actually very similar.

Rob Grafrath (24:18)
Really.

And I'll see you

awesome

Yeah, that's very cool. And LLMs, you will have ads in LLMs soon. I don't know if everyone knows this, but they're already angling to get that marketplace. So have your collection agency. Exactly. Exactly.

Adam Parks (24:42)
You

If you're not paying for the product, you are the product. Yeah, look at I try to remind people all the time that if you're going to feed the LLM, you're feeding the LLM like that's why they're giving it to you for free so that you will feed it with information and examples that it can use going forward. But when you're paying for it, there's a disclaimer at the bottom that literally says we're not going to take the information from your engagements with chat, GPT or whatever and use them to train the model that you're actually segregating yourself. But I always repeat myself that if you're not paying for it, if you're not paying for the product, you are the product.

Rob Grafrath (24:53)
We can have some of that.

Absolutely. Yeah. No, that's, that's, it hits even harder the further into the future you get because people will be so much more information will be commoditized and power information being power. might even find that collection agencies depersonalized information on behavioral metrics could become another revenue stream. You sell information about

Adam Parks (25:23)
I don't think that changes in the future. I think that remains the same.

Rob Grafrath (25:47)
Well, with these messages, we got these kind of contact rates. You don't have to have the consumer information included in that message. That could be very valuable to somebody for you.

Adam Parks (25:56)
Yeah, it'll be signals over intelligence because if you look at the way that signals are assimilated into active intelligence it's going to be about how many of these signals can I collect, interpret and then combine in different ways to learn or drive intelligence for the future. And that is pretty much how like the entire intelligence community works. That's how data science is starting to work and why we're building such big data lakes right now to put as much information in one place the challenge starts to come in with these LLMs in terms of their context windows and how much can the LLM truly understand within any individual response that it's gonna provide. One of my friends recently reminded me that chat GPT is the equivalent of a drunk frat boy. It will speak with the utmost of confidence no matter what the response or its actual knowledge behind the response it's providing.

Rob Grafrath (26:27)
Thank

Cheers.

Thank you. ⁓

And it always wants to please you so bad. Try GPT-5, they say that they've made it less of a complete people pleaser. I haven't played with it enough to tell you whether that's true or not.

Adam Parks (26:57)
I can feel differences between the model. The question is, I always go back to your purpose and using the model because there's some models that are better at other things and you have to use the right models for the right purposes to really get the maximum value and to always try and do things in short bursts so that you can stay within the context window of its understanding in the response.

Rob Grafrath (27:19)
Yeah, prompt engineering. Prompt engineering will be like, you go to school to learn how to program, you go to school to learn math, you're going to go to school to learn prompt engineering. That's going to be great school level. Like, how do you interface with these AIs? We have to give some shout outs here to the, there's some other nerds that, and I'm going to steal one of their ideas, that John Erickson, Heath Morgan and Chris Walcher, we've nerded out in the past talking about this sort of thing and and the idea of if even at the this is again far future stuff so so bear with me but even at the grade school level that everyone will have an ai companion everyone will have their thought so to speak right so this is again sci-fi bear with me but it's running along with It's not that sci fi Heath Morgan and I did a AI hub episode on this.

Rob Grafrath (28:10)
Yeah, it's learning along with you and with the kid and the kids, they're learning how to interface with it, how not to, right? Don't rely on it for the love and nurturing that you need to rely on your human caregivers for, right? Let's make sure people have a healthy relationship with their AIs and with their co-pilots, if you will. But definitely do learn how to use it. If you're not teaching people how to use the things, then they are going to misuse them, they're going to misunderstand them, they're going to, you know, the guys that fall in love with their chatbot because they think that it cares about them more than anyone else in the world because it's the first thing that they've had an interaction with where they weren't judged and treated like, like, like, uh, they're, they're not intelligent or whatever. Like this thing respects them and it's, it's almost sad, right? Some of the stories you hear, but, uh, but this is kind of where education has to come in and has to step in earlier.

Adam Parks (29:06)
Heath and I did an episode talking about how, so I had a daughter a few months back and Heath was talking to me, one about the centralization of even after I pass, there will be a bot that she will be able to engage and interact with. I look, I host a podcast. There's 350 plus hours of video of me out there and I've been able to tie it into a single model and start to play with what does that look and sound like, which is interesting, but.

Rob Grafrath (29:17)
of the body of yourself.

Adam Parks (29:32)
When we were talking about the children and how are the children going to engage or interact with artificial intelligence in the future, I think they're going to be not necessarily graded on the outputs, but graded on their inputs to the models. And how are they actively doing that? And I'm wondering if the education systems are going to be able to keep up at that same pace to be able to teach these things, because think about it from the teacher's perspective.

Adam Parks (29:58)
they now need to learn this whole new world of things that they need to not only become experts in, but now expert to the point in which they can communicate it to third parties and teach others how something functions. I think there's gonna be some pretty significant challenges in that regard and artificial intelligence is going to change education. I was watching something just last night about colleges basically banning chat GPT. And now you have to enter where they're tracking within Google Docs, for example, every keystroke, the timing of those keystrokes, the timing of copying and pasting of content that's coming into this document, and what that ultimately is going to look like over time. And then when you build a model based on that, you should be able to tell within a pretty accurate understanding, right? Did somebody write this?

Rob Grafrath (30:50)
I'm going back to school trying to get a masters and when I there's I'm doing online stuff when people post to the to hey introduce yourself but it's so much GPT. It's just you can sew to style. You got the yeah the paragraph structure that it loves those dashes that that long hyphen that dash loves the little I can tell the GPT just by the way that they lack on something or or a paragraph structure.

Yeah.

Yeah.

I used to use them, now I can't use them anymore.

You

Rob Grafrath (31:19)
I argue, I like to think, is that person that you're preparing to enter into the world, are they going to be living in a world where that tool doesn't exist? Should you maybe teach them, like I Grammarly, let's say, for example, I have Grammarly installed on my computer and it's everywhere that text shows up on the screen and if I'm writing an email or whatever, it's giving me little prompts. You have a typo here, this grammar is a little funny here, just little tweaks to my grammar.

Rob Grafrath (31:47)
I see nothing wrong with that. And if you allow the use of something like GPT to at least generate some content and to polish content, long as it's, I feel like as long as it's used properly and that you can prove that there was human engagement in the content, that that satisfies things. I also think again, going into that far future in education, it's not hard to imagine universities opening up free online classes taught by AI, right? That if it's doing PhD level stuff today, there's no reason they can't teach an economics one-on-one class. And I took statistics recently and the book was like opaque. But I went and had CHPT teach me the subject because I didn't understand something and I would ask it and it would explain it.

Rob Grafrath (32:39)
ask it something else and explain it better. It's your real time, immediate feedback professor. It's a great teacher. And I don't think that banning it or saying that we're not going to let AI teach people is the right road to take. It's probably like, let's make sure that it's teaching people truths, right? And not just the creative stuff that it's coming up with. That might maybe a...

Adam Parks (33:00)
But it's the underlying knowledge, right? So if we only rely on the tools, I'll give you a perfect example. When I was in high school, I started taking accounting classes. And this accounting class was not like how to use QuickBooks. It was physically right, that right up on the board today, today, we're going to cut a check for this, we're going to reconcile these accounts, like they gave us the instructions that we were going to execute for the day. And I remember one of the girls in the class saying, look, my QuickBooks does that, like, I don't need to learn this, this is irrelevant.

Rob Grafrath (33:28)
That's pretty cool.

Adam Parks (33:29)
And the professor's response or the teacher's response was, well, how do you know the QuickBooks is right? Like, how do you know that these things are happening the right way? Like, how do you understand the underlying foundation if you are not understanding what's behind the tools themselves? It's important to learn the tools at a certain point. But prior to that, you have to understand why the tools are doing or what function they're performing so that you can even appreciate the fact that that tool is moving it forward.

Rob Grafrath (33:32)
you

Adam Parks (33:58)
Now, when you talk about the AI teachers, that is a really interesting subject. I'm torn on it. As someone, you know, we plan to homeschool our daughter and in that process, what is that going to mean? Does that mean that there's a physical person traveling around the world with us? Does that mean that there's a combination of a physical person in online live instruction? Does that include artificial? Like, what is that going to start to look like now?

Rob Grafrath (33:59)
other things that we can try to feature as we go. So,in a company and a single person.

Thanks for listening.

Adam Parks (34:22)
I've been learning Portuguese through Duolingo for a couple of years now, and I can tell you I don't speak Portuguese. I mean, a little bit, but I don't know how much I can actually get from that device to really put me in a fluent position to communicate.

Rob Grafrath (34:28)
Thank

Yeah, but I do believe, let's say Duolingo gets a AI powered version and you can have a conversation with it in Portuguese and then have it correct you like your pronunciation on that was a little off. Let's try this. I just feel like because yeah, it's it's you can't replace the greatness of a human teacher, but you can clone it, right? You can reproduce it in mass.

Adam Parks (34:43)
The rest.

Rob Grafrath (35:03)
And then let's say you return to class or you go back to the real world and now the human says, you've done great at learning this. You need improvement here. So just kind of like, it's almost like the human becomes guardrails, I feel like. that hopefully curates and guides the AI teachers of the future. mean, you're going to have AI teaching you how to interface with AI. It's so meta.

Adam Parks (35:04)
Make it available. That could be the case. That I think sounds more realistic these days. Well, it's the judge, it's the LLMs watching the other LLMs watching the other LLMs. And as you start thinking about these judge LLMs and how everything interacts with each other, it really starts to become a little bit clearer in terms of what our future is going to look like as an industry.

Rob Grafrath (35:33)
Yeah. of the cell really starts to become a computer. LLM's all the way down guys. Kinda scary.

Adam Parks (35:47)
It really, and what is a large language model if not a giant math problem? And so for me, as I've gotten deeper and deeper into this, I believe that the order of operations is the core of math. I believe that the order of operations is going to be the core of our prompt engineering, at least in the near to midterm. And then who knows what kind of neural links we come up with in the future and how all that will play out in the matrix. But that might be a different story for a different day.

Rob Grafrath (36:16)
You're right on.

Adam Parks (36:17)
Rob, this has been an absolutely fantastic conversation. I really do appreciate you coming on with me today, sharing your insights and really just speculating on what this future might look like and to hear it from someone who's got such a creative futuristic mind frame has been really interesting for me.

Rob Grafrath (36:20)
Thank you.

Thank you. I want to end on a quote from Yogi Berra that the future ain't what it used to be.

Adam Parks (36:41)
Okay.

I like that's a good one. The future ain't what it used to be. I think that is the perfect quote to end on.

Rob Grafrath (36:48)
Ain't what it used to be.

Adam Parks (36:50)
For those of you that are watching, if you have additional questions you'd like to ask Rob or myself, you can leave those in the comments on LinkedIn and YouTube and we'll be responding. Or if you have additional topics you'd like to see us discuss or things you'd like to see us speculate wildly about in the future, you can just leave those in the comments below and hopefully I can get Rob back at least one more time to help me continue to create great content for a great industry. But Rob, thank you so much for coming on and sharing your insights today.

Rob Grafrath (37:11)
Thank you so much.

Thanks for having me, Adam.

Adam Parks (37:16)
Absolutely, and thank you everybody for watching today. We appreciate your time and attention. We'll see

you all again soon. Bye