With all its associated technologies (neural networks, symbolic reasoning, search algorithms, probabilistic reasoning, expert systems, and more), AI has been evolving in universities, the government, and corporations for decades. But only those with a keen interest in this technology have been paying close attention to its progress. Meanwhile, a chatbot helped you when you contacted support. Or you used Alexa or Siri to answer a question, play some music, turn on your lights, buy something online, pay your bills, or tell a joke. In our lives, we have caught glimpses of AI’s potential but no clear line of sight as to how powerful the underpinning technologies have become or how quickly they are evolving.

With the unveiling of ChatGPT and similar tools, we are now face to face with the AI era. In his book, Impromptu: Amplifying Our Humanity Through AI, Reid Hoffman says, “Much of what we do as modern people—at work and beyond—is to process information and generate action. GPT-4 will massively speed your ability to do these things, and with greater breadth and scope. Within a few years, this copilot will fall somewhere between useful and essential to most professionals and many other sorts of workers. Without GPT-4, they’ll be slower, less comprehensive, and working at a great disadvantage.

The Payments Industry is interesting because it is constantly changing, and technology has always been a significant change agent. For decades, networks, processors, PSPs, merchants, and financial institutions have invested in technology to increase adoption, create new services, manage risk, and accelerate initiation, clearing, and settlement. 

In this episode, Yvette Bohanan is joined by Frank Young, a 35-year veteran of the fintech space, and Glenbrook’s Russ Jones to think about how AI might transform the payments industry and how organizations should mobilize for this transformation. It’s time to start talking about this topic seriously – with each other – and not just asking ChatGPT. 

Yvette Bohanan:

Welcome to Payments on Fire, a podcast from Glenbrook Partners about the payments industry, how it works, and trends and its evolution. Hello, I’m Yvette Bohanan, a partner at Glenbrook and your host for Payments on Fire. It’s the summer of 2023. Over the past year we have emerged from a global pandemic thinking we were navigating into a new normal that was quite comfortably actually based on a fairly well-established old normal, with a few improvements. Our thinking abruptly changed with ChatGPT and the big reveal that is neural language modeling. We didn’t expect this new normal, but we should have, because ChatGPT and tools like it give us direct access to AI technologies that have been gradually creeping into our lives for years. In his book Impromptu: Amplifying Our Humanity Through AI. Reid Hoffman says…

Recorded Clip – Hoffman, Reid. Impromptu: Amplifying Our Humanity Through AI (pp. 5-6). Dallepedia LLC. Kindle Edition.

Much of what we do as modern people at work and beyond is to process information and generate action. GPT-4 will massively speed your ability to do these things. And with greater breadth and scope within a few years, this co-pilot will fall somewhere between useful and essential to most professionals and many other sorts of workers. Without GPT-4, they’ll be slower, less comprehensive, and working at a great disadvantage.

Yvette Bohanan:

That’s a lot to consider. So we’re going to start small on this episode and just look at the possibilities AI presents to the payments industry. And to embark with me on this conversation that will likely continue for many years to come are Russ Jones, a partner at Glenbrook who leads our education program, and our extraordinary guest, Frank Young, a 35-year veteran in the FinTech space who we’re catching up with during his hiatus after a successful eight-year run on the executive leadership team at Global Payments. Frank, Russ, welcome to Payments on Fire. It’s wonderful to have you on the show.

Frank Young:

Thank you. Thank you very much.

Russ Jones:

And Yvette, it’s always great to be on Payments on Fire, one of my favorite podcasts.

Yvette Bohanan:

Your biased opinion here, Russ. Favorite podcast.

Russ Jones:

An extraordinary podcast about strategic issues in the payments industry.

Yvette Bohanan:

With extraordinary guests. Frank, it is so great to see you and to have you on the show. Before we dive into our topic at hand and why we’re here, I want to make sure that our listeners get the advantage of listening a little bit to you describe your career because it’s one of the most fascinating careers in the industry as far as I’m concerned. Can you share a little bit about what led you into payments to begin with and what inspired you to pursue specific opportunities?

Frank Young:

Thanks, Yvette. Who doesn’t like talking about themselves? So my FinTech career was born in 1988 and I went on to ChatGPT and asked it, who’s the most famous person born in 1988? And it told me that Adele, the singer, was born in 1988. So my FinTech career is as long as Adele is old. I do not sing anything like her, but I’m pretty sure she couldn’t spell many of the acronyms in payments that I can spell. I’ve been at it since 1988. What got me into it was I was part of a management training program at Chase Manhattan Bank many years ago, and they gave us a rotational assignment where I was fortunate to be placed in the retail bank payment strategy group of Chase Manhattan Bank at the time.

And I was responsible for two products. One was a telephone bill pay product, which actually worked on rotary phones. The listeners may not even know what a rotary phone is.

Yvette Bohanan:

Wait a minute, telephone bill payment that you could use with a rotary dial phone?

Frank Young:

With a rotary phone, yep.

Yvette Bohanan:

I had no idea this was even possible.

Frank Young:

We actually convinced customers in 1988 to pay us $4.95 a month for the privilege of using their phones to pay bills. It was actually pretty successful.

Russ Jones:

And the value proposition was all about eliminating stamps.

Frank Young:

Stamps, that is exactly right.

Yvette Bohanan:

Impressively phone bills in ’88 were usually under $100, so you didn’t have to put in more than four figures.

Russ Jones:

I can smell the value proposition a mile away.

Frank Young:

I was responsible for that product and I was also the bank’s representative on the board of at the time, the NYCE ATM network, which was partially owned by Chase and other banks. And so that was my entry into payments. And interestingly enough, doing that work, I had gotten my undergraduate degree. I enjoyed it so much. I remember asking the leadership and my mentor at the time, if I wanted to continue to do this kind of work, what’s the career path? And they recommended business school, which I hadn’t really contemplated before that, and was fortunate to finish my degree program at Wharton. And from there went into consulting, spent a lot of time consulting, issuing banks primarily on the issuing side.

At some point I was considered Accenture’s global subject matter expert on the topic of risk management, which had given you a sense of where that had taken off. And then I moved into a position with MasterCard at the start of their consulting business. Got so passionate about mobile technology that I joined a startup in Atlanta called Firethorn, which was acquired by Qualcomm. And what we built at Qualcomm with Firethorn was exactly what Google was planning to build with Google Wallet. And so I parlayed that role into a position on the West Coast and moved my family out to Google where, Yvette, you and I had the pleasure of battling many demons and pivots and strategic changes to digital payments. And after Google, I landed in my most recent job at Global Payments.

And the interesting thing is that 35 year career really gives me the benefit of a number of significant epochs in payments that allow me to see some patterns. And so when I was at Accenture Andersen Consulting, it was at the dawn of the internet. And I remember talking and being passionate about what the possibilities were. And I remember hearing from a partner around the fact that our banking clients have no interest in the internet because all it is is “bloated brochure ware”, was how they described the internet. And I was dumbfounded that that was the position because I saw something much, much more significant. Fast-forward to the dawn of the mobile web working at Firethorn, which was very much a carrier centric solution.

And Apple was coming out with the iPhone and I was trying to encourage a more open architecture to what we had built at the dawn of the mobile web, not be so beholden to the carriers. And many of the people leadership at Qualcomm were, “well, the iPhone’s going to be a flaming disaster”, which wasn’t the case obviously, pretty much turned our entire business model upside down. And so what I saw at the dawn of the internet and the dawn of the mobile web, I like to think I was an early adopter of cryptocurrency. I bought my first Bitcoin in 2013, which wasn’t really, really early, but it was pretty early. Saw the promise there. And then I look at the time period we’re in right now with the emergence of large language models, artificial intelligence, people are increasing the chatter about how much closer we’re coming to AGI, which is quite different from AI and I’m not really here to talk about that.

But the exponential times that we’re living in, when you take Web 3.0 concepts, cryptographic developments, the blockchain technology and artificial intelligence, I have never been more excited about technology. And I’ve just given you a sense of some of the other major waves of change that I’ve experienced firsthand in my career. I think one that we’re in right now is unlike any other in terms of the impact it’s going to have on society and humanity. I feel like I’ve won the lottery that I’m alive to witness it and be in the business world to help participate in it.

Yvette Bohanan:

So Frank, there’s two firsts happening right now. The first one is you are the first guest to show up on this podcast and throw shade at Adele. And so I’m just going to, full stop. Perhaps the only guest that’ll ever come on, but you never know. Second is this is the first conversation we are having on the podcast about specifically the current moment we’re in, the current era is that you’re starting with this notion of AI front and center in things. And that’s what we’re really talking about. Now, this is not going to be the last time we talk about this, but what really grabbed my attention was, I watch what you post on LinkedIn. I watch a lot of people. And you wrote an article and I said, after I read it, I was like, I wish I wrote that.

I wish I had had the presence of mine to put that together, put it down on paper and publish it. So thank you. And I guess I liked it so much because it’s exactly the way I was thinking about things. I’ve been reading a lot since ChatGPT4 launched and everything. Everyone’s been coming out and saying, how is this going to affect payments? Everybody’s been posting about it. People are writing these eBooks and things about it already. And my thought was, you’re not thinking about this hard enough. You’re not thinking about it deeply enough. You’re not actually absorbing what this could be. And I posited to some people around Glenbrook, we have this thing in our education program where we have the eras of cards and payments and innovation and that.

And we’ve been saying that since about 2009 we’re in this digital era. Did I get that right, Russ, roughly?

Russ Jones:

And we are.

Yvette Bohanan:

And we are. But we kept saying, well, in the 2020s something’s going to happen here. I think we’re broaching at the precipice of what I might call the AI era is the next. And that’s what I wanted to talk with you about. So can you summarize what you said in the article for everyone, because I won’t do it justice.

Frank Young:

Thanks. I think I’m a big fan of, whenever I’m faced with a complex problem or a curious question, I always say to myself, do you really understand the nature of the thing you’re dealing with? What is the game that’s being played and what is the nature of it? And I trying to get back to first principles. And so I think I’m actually in this time off, I’m going through a certificate class with MIT on blockchain, and some of the early content that they talk about is what is a general purpose technology? And they actually propose the suggestion that blockchain is not a general purpose technology, because what makes something a general purpose technology is, one, it’s pervasiveness. Where will it show up? Where will it be applied? Who will have access to it? How widely dispersed will the technology potentially be?

And I think you’d be hard-pressed to say that aside from the people who are passionate about blockchain and people who are studying it, it’s not showing up everywhere. You know can have a cocktail party and not many people will be able to really have a deep conversation about a blockchain. But if you ask them about this cool thing called OpenAI, it’s the kind of thing I could take out of my phone and show my mother-in-law, God bless her, who’s 80 years old, and explain to her what it does and the general gist of what’s out there. So it’s pervasive. The scope of a general purpose technology also has the characteristic that its impact will be broadly applied. And so AI blockchain may have been very relevant to FinTech as an example, but does blockchain really have an application to pick an industry, automotive, I’m sure there are unique applications for it, but maybe not as applicable as other industries.

So it’s pervasive. It needs to have a broad scope of impact. It will undergo a series of continuous and rapid improvements. And I don’t know any other area you’re seeing more continuous improvements than in AI. AI with the release of 3.5 to four was a dramatic improvement. Some of the things you’re seeing in the applications like Midjourney that are being developed, not a week goes by where my mind isn’t blown by what can be developed through text to image, text to video, images to video, everything that’s happening there, the pace of continuous improvement. And then finally it spawns further innovation. And so when you look at the AI, it checks all those boxes, unlike blockchain, unlike quantum computing, unlike some of these other things we’re hearing about.

But the other things that check those boxes is electricity, is the steam engine, is the internet generally, as a general purpose technology. These are massively innovative capabilities that force you to think that this is going to be put into the hands of everybody. And back to pattern recognition, I remember at the dawn of the mobile web, every major company defined this new role within their company, head of mobile strategy, head of mobile banking. And they may still exist because they’re the product manager for a mobile application, but I think most people today would recognize there’s no need for somebody to only focus on mobile strategy. I mean, it’s-

Russ Jones:

Frank, what that reminds me of, in the early days of the internet, in the banking industry, online banking was different than banking. It was viewed as a different discipline and banks would sign people up to be a customer and then they’d start over again to sign them up to be an online customer.

Yvette Bohanan:

Right, exactly. Exactly.

Russ Jones:

And I think your point’s excellent. That’s the same thing going on with mobile. It’s like it’s not enough to be a customer, you have to be an online customer and now you have to be a mobile customer on top of that.

Frank Young:

I agree. And sometimes I try to figure out what would life have been like in the early 20th century at these industrial facilities, was somebody determined to be the head of electrical production versus the head of non-electrical? At some point it just became everything, everywhere.

Yvette Bohanan:

It’s how you do. It’s how you do it.

Frank Young:

That is what AI is. And I think that my thesis is, there’s real risk for you delaying in that recognition. You won’t make the necessary investments, you won’t do it at a scale that’s going to capture the full benefits of it. You’re going to be subject to some significant competitive pressures and that the sooner you get there, the sooner you recognize the nature of the thing you’re dealing with, the better off you’ll be positioned competitively, both at the company level, but at what I tried to stress in that article because it was LinkedIn, just in your own personal career management. Immerse yourself in this because this is like somebody invented the Sawzall when you’re a carpenter and you don’t know how to use the Sawzall.

How many more homes are you going to be building if you don’t take those indispensable tools and apply them to your craft and try to build your career around it? So that’s really the central theme, is you define it the right way, understand the implications and get going, because it’s, act accordingly I think is even in the headline, because I think a lot of my own observations are that there’s still a lot of people in corporate America who are like, yeah, it’s a nice toy and it could generate pretty pictures, but I’ve got numbers to hit next quarter and I really can’t devote the time.

Yvette Bohanan:

I can’t mess around with this stuff. This is a distraction. This is a rabbit hole.

Russ Jones:

I was smiling when you’re talking about the impact, maybe it’s not as much on the automotive industry. The counter argument would be that it’s your car that’s the first introduction to AI for most people these days. I think about my Tesla, it can see, it can talk, it can think, it can make recommendations. It’s amazing in an AI sense.

Yvette Bohanan:

Just to touch on that a little bit more, when we were working together at Google, Frank, work was underway, well underway in this arena. When I was back at Vesta and we were very early days of predictive modeling. All of these things have been building up for a long, long time. I think it was this notion that it’s accessible. What ChatGPT did, what Sam Altman did, is he basically blew the doors off the thing and said, this has been under wraps and it has been the domain, if you will, of a few. Like blockchain, it’s the domain of a few, it could be thousands, but it’s still a few compared to global population. And now it’s accessible to everyone. You can take it out, you can show it to your mother-in-law, you can show it to your kids, your kids can play with dolly, your kids can do this, your kids can do that.

And all of the sudden people are trying to get it to write sonnets or something. Who even sits down with a pencil and paper and tries to write a sonnet? But that’s the first thing a lot of people went to. I want to see how clever this is. I want to test it.

Russ Jones:

One of the first rules in business of that is the thing that you can see always wins out over the thing you cannot see.

Yvette Bohanan:

Exactly, exactly. And here we are. And so let’s go back to your point then, because this is where I wanted to get us in this conversation. I want to go a little bit deeper with you on this notion of, because you did start your article out with a nod to Tracy Wilke and his post. And which I thought was great. And bringing it home to people, if you’re not doing this, you need to be right now. And after I read your article, I was listening to a podcast by Reid Hoffman and Sam Altman discussing ChatGPT, discussing Reid’s book, Impromptu, which is like a travel log exploring all of these things. It is great book. And they said something that resonated. So I want to play a 40-second clip and then I want to explore this with you a little bit more.

Recorded Clip – Reid Hoffman and Sam Altman, AI Field Notes, Greylock

I think the greatest platforms people pretty quickly forget are platforms. It was like a while after the first smartphones and app stores launched and all these companies were saying, I’m a mobile company. And that was a big deal, because they were on the mobile platform. And today it’d be ridiculous to say you’re a mobile company because every company has a mobile part of its strategy. And we’re going through a moment right now where everyone’s talking about AI. AI is talking in the future, probably true, all these AI companies, whatever. My hope is that 10 years from now, intelligence is expected in every product and service and it’s so ubiquitous, we forget it’s a platform. It’s just part of everything we do.

Yvette Bohanan:

So let’s dig into the how-tos. Okay? What should companies be doing or not doing to get smart about what’s happening and get into the game and not just consider this that side distraction. What’s the step? What’s the next step?

Frank Young:

Well, I’ve got a view, and I come from eight years of working at an organization who was very much execution focused, to a great degree to their benefit because they were very predictable. They understood the concept of setting targets and hitting those targets or exceeding them and doing that consistently. And a lot of publicly traded companies get hit with the fact that they’re too short-term focused and that they focus there. But there’s a role for that level of discipline and a lot of what happens in large organizations get set at the agenda of the executive team, who then drive the entire organization. And there have been a few instances over the last several years, both at Google, at Global Payments, other organizations I’ve worked with, where at the very highest levels in the organizations, they’ve set those goals and objectives. They’ve set the direction they want to take the company.

And so we had the benefit of being at Google when Larry set the mobile first strategy. And for a year every product organization within Google had to reconfigure or design their products and services to work in a mobile only world. I think there was one day a week where they were saying, come to work and leave your laptops at home, work off of a mobile device only, just to get used to living in that environment. There was the year of social, Google Plus every product had to incorporate Google Plus. When I was at Global Payments, we had a year where we all had to focus on deploying multi-factor authentication in every customer facing application we had. And everybody’s roadmaps got turned upside do-wn because so much of the development activity was turned toward deploying multi-factor authentication. And so those kinds of initiatives tend to focus the organizational mind.

So number one is, I would encourage executives and boards to really consider establishing a mandate to simply state AI everywhere, put AI in every aspect of the business. And there are ways a lot of companies can do this. Most companies, I know we had this at Google, anytime we wanted to surface a business case, we would go through a committee, we’d put a short presentation together and we’d explain what the customer value proposition was, how we wanted to see this executed, what kind of roadmap it would be on. And then there would be the investment ask. I think it would be very simple for companies who have that kind of process in place, to simply make the requirement that every business case over a certain dollar amount has to explain how it’s going to leverage AI to achieve its objectives.

Now the AI could be a part of the output, it could be a part of the product that’s being developed or it could be a part of the tools you use to build whatever it is you’re working on. So we’re going to use AI to support our testing methodology at the conclusion of our development effort, and that’s going to allow us to shorten the testing time from X to Y and clarify the value proposition of why AI is going to matter. But I think that has to be a top-down directive, somehow built into the processes the companies have to make sure that it’s getting deployed. And that’s a top-down approach. I think another element of the top-down approach is you’ve got to put some guardrails around it. There have been some stories of developers uploading proprietary code to ChatGPT to optimize it, and given the IP rights, does that then become a part of their data set? Who owns it there?

You’ve got to be careful. You can’t just indiscriminately roll this out. But there could be an element of organizations encouraging people in a lot of ways to start using the tools. I love Reid Hoffman’s use of the word co-pilot. I also like when people refer to AI not as artificial intelligence, but augmented intelligence. Think as an employee of the company you work for or as just an individual trying to chart their way through the world, you now have an ability to always have at your hip, literally the smartest person in the world to ask any question to. There’s a point of view out there that suggests that the true innovation of ChatGPT, is that it has hacked the human operating system, the human operating system being our language. The only way that humans progress the way that we do is because we can communicate through language.

And the large language model has in essence captured all of that and put it into a usable form. And so again, top down is build it into the processes, set the requirements and mandates, the guardrails around how you want to do it and look for ways that you can encourage the use of it. As individuals, me personally, when I’m working at home, I’ll have two browser windows open on my desktop pretty much 80% of the day. One of the- browser windows is logged into ChatGPT. And one is what I’m doing through the day. And what I find myself doing fairly constantly is going back and forth between my normal browsing, my normal research, my normal work, and going over to ChatGPT and feeding it some information. Now what I’m finding is it’s not always correct with ChatGPT. I’ve tested it a few times around, I think I asked recently, there’s a big debate in Atlanta around whether or not Atlanta should build a second airport.

And I was wondering how many other big cities only have one airport? I asked ChatGPT, build me a table, laid out like this, the top 15 MSAs, the airports within 50 miles of the center of the MSA. And it came back and told me that three cities had at least more than one airport. And I said that that’s not right. I know Dallas as an example has DFW and Love Fields. And that was a mistake. I went in and said, I think you made a mistake. What about Dallas? And it came back and said, sorry, you’re right. And it added it and I found another one and it said sorry.

Yvette Bohanan:

Right? Yeah, I’ll update that.

Frank Young:

I found like four errors in it. So it’s not always correct. You can’t always just trust-

Russ Jones:

It’s only as good as the data that’s behind it.

Yvette Bohanan:

Exactly. It’s still predictive modeling. It’s worth mentioning, AI is a conglomeration, that’s not the technical term, but it is a conglomeration of tools and technologies. And a lot of the technologies are predictive. A lot of them are knowledge based. So you have hard facts and it’s just basically got the facts and it spits it out. But what we’re seeing now is this opening up. It’s like we’re at a pivot point with the neural language aspect to it of the model. I like what you’re saying, it opens up this human operating system. We can now start to explore this with more depth than going to your favorite search engine and just typing in the toolbar some complicated query. I think that’s the difference, but it hallucinates, right?

Hallucinations are there. It will put things together and sound really good, but it’s not quite right or it’s dead wrong or it’s missing information, because, to your point Russ, it’s only as good as what’s in there. And I’ve hit a couple things myself right now where ChatGPT says, I don’t have any information for you because the data that I’m working with stopped at 2021. So there is this notion that it’s a co-pilot to you, but you’re a co-pilot to it. You’re literally co-pilots flying the craft or whatever, when you’re doing this.

Russ Jones:

My observation here of that, I love, Frank, you were talking about your early days at Firethorn. And in that period of time, every company in the world wanted to have a mobile strategy. And I always could imagine in the so-called C-suite, the CEO saying, okay, you knuckleheads, where’s my mobile app? And every bank wanted to have a mobile app. Every enterprise wanted to have a mobile app. And what they found out was you couldn’t have a mobile app until you had an API strategy. And I think the same thing is going to happen here with AI. You can’t have a rational AI strategy unless you have a rational data strategy.

Yvette Bohanan:

Absolutely.

Russ Jones:

And you really get your arms around the proprietary data you have, how you’re going to leverage it, how you’re going to keep it pristine, manage the integrity of your data. And if you don’t have that as an asset to work with, you’re going to have to have a great strategy on where your data’s going to come from, because you’re not going to get the results you want if you’re using the wisdom of the crowd when the crowd’s half paying attention.

Frank Young:

That’s another great lead-in to another point I was going to make around the top down approach, is I think from a strategy perspective, if there was somebody who was designated and responsible for AI, I think the question they would want to ask themselves is, where in my organization do proprietary data sets exist and do they provide us a knowledge graph that we can build our own large language models on top of, that are unique and tailored to us? Because I think the option value of that is if you can build a good enough one, not only can you use it for yourself, but then you can expose it to others to potentially use to take a little bit of the open-minded approach. And not everybody has this enlightened view. They think that I have a proprietary dataset I can build a knowledge graph off of and create my own proprietary LLM. I’m going to hoard it to myself and only use it for the benefit of myself.

Well there’s one school of thought that says, well why not open it up? Give it away to everybody and you become in essence almost, take Tesla for example, if it creates the greatest autonomous driving AI capability, what would benefit them and humanity the greatest by them only using it for Tesla vehicles or figuring out a model where they can expose it to others to allow them to incorporate it and almost become a platform if you will. So that second part of opening it up and becoming a platform for others is secondary to just asking the question, where do these knowledge graphs exist? And FinTech players, look at your service organization. Do you even have the ability to listen to what your customers are calling and complaining about and how you are resolving those queries?

And if that became a knowledge graph and you became the best at, I’m sure the questions that one FinTech isn’t that different from another FinTech, my check didn’t clear, that transaction failed, I got a chargeback over here, you misreported this or that. That can become an immensely powerful opportunity for somebody to say, and Russ to your point, do I even have the data infrastructure to store in a privacy controlled way, voice recordings bereft of personal information, but taking it and learning from the questions they’re asking, the resolutions we’re providing and honing that over time? And that’s just one example. I would imagine companies have 15 examples of that somewhere under their hood of how they’re handling things and every function in an organization can contribute to it.

The accounts payable department, are you listening and reading the emails of the correspondence with people you’re paying around delayed payments and how are you managing that and could you potentially turn that into a more effective and efficient way to service vendor relationships? There’s not an organization within a company that won’t have that opportunity, but Russ, you’re spot on. Somebody has to get the data in a format and an environment where you can start to build those assets around it. That’s an enormous opportunity and challenge.

Yvette Bohanan:

And we have always said it in payments in particular, manual processes and processes that rely on one person to do something all the time, and there’s a lot of mundane stuff in payments. The payment operations isn’t all unicorns and rainbows. There’s like recon and there’s all kinds of stuff that goes on, that oftentimes it’s one person sitting there with a spreadsheet still today dealing with compliance, risk management, reconciliation and financial operations, vendor management, contract management with your suppliers. I could go on and on and on. Tons and tons of companies are still doing all this stuff manually. Chargebacks.

And if you’re still doing it manually, maybe the first thing to ask those teams when you put this AI strategy, we call it a BHAG, right? The big hairy audacious goal out there, is when you’re writing up your particular objectives on a quarterly basis or whatever you do, how am I getting rid of the manual process? How am I starting to curate my data? How can I incorporate the technology to help me do that? How can I incorporate the technology to build off of that? Depending on where you are in that spectrum of automation is going to determine where you have to really start. But you have to get started. If you identify a ton of manual processes right now in your organization around this space, dig there first, start there first.

Frank Young:

And then that calls in maybe a little lower level than setting the goals and the objectives and building in the processes. I think there’s some fun things companies can do. I don’t know if the term’s been used before, but if not, I’ll trademark it. Everybody’s familiar with hackathons, but we need to do hack-AI-athons. Give people tools and say, come up with a way to do your job better by using this augmented intelligence in the form of an AI tool. And now we’ve got to give them the tooling to do that. But I think if you give people the ability and rewards and recognition and prizes and spiffs and things of that nature, I think you could push this down to every corner of an organization and get some excitement about finding applications for AI.

And that concept, applied use of AI, is to me in my break here in between careers, the applied use of AI in a business is a really, really deep pool that I’m trying to figure out how to turn it into a career opportunity for myself. And I’m sure others are thinking along the same way.

Yvette Bohanan:

Well sure, but you know what, you’re bringing up something really interesting and it’s reminding me of a few things. First of all, it should be fun. You can have a lot of fun with these tools right now. The second is, when I was at BofA, working on payment strategy on the consumer side of the bank, we took a little bit of a chapter and verse out of what we did when I was at Amazon, which was you write a press release. And it’s a one-page press release that explains what you’re going to do and why it’s cool and whatever, the key aspects of the idea. And we put people in a room, we put about, I think we had 70 people, we had 10 groups of seven or seven groups of 10, and it was cross-functional, deliberately cross-functional representation. So we had attorneys, we had legal team members, finance members, all sorts of associates from across the bank.

And they basically were paired up and we gave them prompts, like “how might we” prompts that we used at Google, but a little bit different, about what they could do. We came up with 17 patent ideas that we turned in after a two-day session. You know can do stuff like that, and it really gets people energized. And I think a lot of the press that we are hearing right now is people are getting afraid. They’re getting afraid that this efficiency’s going to show up at some point and they’re going to be out of work. Their job’s going to be eliminated. There’s going to be no need for humans, robots are going to take over the world, dah, dah, dah, dah.

Russ Jones:

I just saw a cartoon last week that we’re where one person says they’re afraid that the robots are going to take their job. The other person says, trust me, the robots don’t want your job.

Frank Young:

I had the observation about the fact that this is one of the technologies that actually is perfectly aligned toward disrupting. I try not to use that term too much, but disrupting white collar work, right? Going after white collar work. But I tend to have a more optimistic view of technology in general. People don’t get pushed to the sideline, but the nature of work changes. The things you’re focused on. And my hope is to the degree we can focus the AI on the lower value tasks, it gives humans the ability to focus on different value or higher value activities. We’ll see what plays out at the end. That notion of when do we start seeing the margin improvement with the real productivity gains? I think we’re still too early to see that.

We’re not going to see the wholesale shutting down of service centers with the replacement of AI agents. Maybe you’ll see examples of quick serve restaurants being completely automated and bots cooking your burgers and chicken sandwiches and pouring your coffee, but I don’t think you’re going to see the widespread application of that for some period of time. But the stores themselves will become much more efficient, because instead of having four full-time staff, maybe you can do it with three plus an augmented agent of some sort that helps you listen to orders and place orders and things of that nature.

Yvette Bohanan:

I think this notion of “how might we”, is a really powerful questioning tool to just examine every aspect. Don’t go through your day just making the donuts, same thing day after day after day. Think about how might we do this differently? How might we do it better? How might we advance so we can have higher and best use of our time?

Frank Young:

It’s really interesting when I’ve played around with the tool, OpenAI in particular, with ChatGPT, one of the unintended results of my interaction was how it helped me ask better questions. And so even encouraging your people to get good at prompting, get good at asking questions, because a lot of people, I say even in conversation sometimes, ask a closed-ended question, get a closed-ended answer. Do you believe AI will be big? Yes. Is different than saying, explain to me the reasons why you think AI will be big and getting a more fulsome answer. Now the AI is going to give you the answer your question asks for, but it forces you to really think through the way you’re asking the question. And there have been a lot of people who’ve in a very short period of time crack the code on getting good answers from ChatGPT, by way of telling the agent what role you wanted to play, right?

Simply by starting your question as, I want you to act as my trusted investment advisor and give me advice on how best I should think about allocating my tax refund in a portfolio of securities including bonds, ETFs. And it’ll give you a very fulsome answer, but it required you to really tailor your question. I think that’s enormously beneficial. If we can just train people in our organizations to have what I call the scouts mindset, to be a scout, to not just simply take an answer on its surface, but really dig, frame your questions more appropriately, learn how to generate the prompts, get good at it. I think that the benefit of that could have far and wide-reaching implications for just how we interact in our respective organizations.

Yvette Bohanan:

I think curiosity is our friend here, just like any time you’re faced with something new, don’t run a retreat, be curious. I think at an organizational level, being curious, not being a leadership team that’s saying that’s a distraction, stop playing with it. But being curious and saying, what are you doing with it? Why are you so fascinated by it? What do you think we could do differently here? Putting those feelers out and understanding why you might want to be making that BHAG, that big mandate about AI everywhere, is a good first step. Russ, we often talk about in our workshops the different ways you can think about payments, right? We talk about it at a systems level, we talk about it as a payment domain level. We talk about the stakeholder level.

Where should people start when they’re thinking about this and applying it to payments? What part of the framework would you dig into given everything we’ve just described about the technology?

Russ Jones:

There’s a lot of things you can do in our assessment. A lot of things you can do with AI technology and the world of payments. Probably the low hanging fruit is you can make better, quicker decisions and probably better decisions. So anytime you have decision points in a transaction flow, and it’s not just the transaction flow, a relationship, it’s decisions about signing up and vetting partners, evaluating another company’s credit worthiness, evaluating whether or not a transaction is likely being done without authorization of the account holder, where you’re making a decision is really the low hanging fruit. And I think that cuts across all payment systems. I think it cuts across the domains, certainly cuts across country borders and it probably from a stakeholder point of view, it probably doesn’t apply as heavily to consumers as it does to certainly the providers of payment services and probably merchants to some degree.

Yvette Bohanan:

I agree with you. I think the other place that’s really interesting is just those mundane things. Like you said, the robot doesn’t want your job. There’s a lot of stuff in payments, engineers and product managers are like, I just don’t want to deal with this, we don’t want to open this up. We don’t know what we’re going to find. We’re sitting on tons of legacy code in this industry. I don’t know if it goes back as far as rotary dial bill pay, but it goes back pretty far still and it holds people back from doing things. We were talking about this, Frank, the other day, implementing fast payment system technology can be really hard, because you’re sitting on legacy code, because your counterparties can’t deal with it, because these systems don’t have the capability to handle API calls right now or whatever it is.

And we had Matera on in Q1 talking about how much legacy refactoring or legacy replacement they had to do in Brazil to get Pix up and running. We’re facing the same thing here. We have three different systems we’re trying to implement now, and that’s just one system. I think there’s huge potential in engineering to say, what can this help us with in terms of getting out of the legacy conundrum too.

Russ Jones:

The final point I wanted to make about the impact on the payments industry is not every stakeholder is going to be advantaged in an equal way. And the holy grail for a long time has been volume, right?

Yvette Bohanan:

Oh yeah. How many items, how many of this, how many of that.

Russ Jones:

How many items do I handle? How many transactions do I authorize? How much money do I move? That type of thing. I would say, the more of the things you do, the more advantaged you are. So a business that in real mundane terms, would you rather have five strategic customers or would you rather have five million non-strategic customers? And I think the companies who have more transactions are going to be advantaged over companies who have seen less transactions. So in some ways, transaction count matters more than transaction amount, and the more counterparties you deal with, the more advantage, the bigger your data sets will be, the more if you’re running a marketplace, the more people that are in your marketplace, all that type of stuff is going to advantage those types of companies.

Yvette Bohanan:

And this certainly can help people scale and scale faster. That’s the idea here. The economy of it and the efficiency of it is going to take over. So last question here before we wrap up, given that we could go on talking about this for another three hours probably, are we in the AI era? Are we going to claim it yet?

Frank Young:

I’ll suggest, I think we’re in it, but I’ll say we’re probably one out in the top of the first inning of the best of seven series, which means we’re very early. So we got a long way to go. There’s going to be room to pivot. And even in those instances I mentioned at the start where these waves came and people poo-pooed certain technologies or didn’t move fast enough, I think there’s always room to course correct and catch up and let others take the slings and the arrows. The pioneers always take the arrows, right? That you may not be lost if you let some others chart the way and delay a little bit. But you’ve got to be aware that that’s what you’re doing. And I think some people are doing that and they’re aware they’re doing it, but they’re ready to act when the time is right.

There are some people who are sitting back because they haven’t yet painted this with the right brush. And through this podcast and other interactions I have in the industry, I hope to shake a few people awake to say, listen, this is coming and it’s coming fast. And there will be riches for people who are able to take advantage of it. It may not put you out of business if you don’t act fast, but you’re leaving a lot on the table by not taking action now and figuring out a way to put this into the hands of everybody in your organization to help you improve how you do business, how you treat your customers, the value you deliver in the marketplace.

I really hope boards and executive teams start setting some of those big audacious goals for their organizations and start employing this technology that is truly magical and powerful when applied in the right way.

Russ Jones:

Here’s a counter view about how this has sprung on us out of nowhere. I was really excited about all this AI stuff, particularly the payments a couple of years ago when I found out that the core algorithm inside of Google Maps driving directions and propagated across any other app that offers shortest pass algorithms, that core algorithm was developed in 1956. And I thought, oh man, you’re telling me we’ve been working on this for 65 years, 67 years, and what’s new? The basic shortest path algorithm about how to get from point A to point B. We’ve known how to do that forever. What’s different these days is how fast we can do that, how many external factors come in to making, that can influence what that shortest path is.

And sometimes I wonder, and I really, I’m serious when I say this, I wonder if we’re doing, Frank’s point was maybe it shouldn’t be called artificial intelligence, maybe it should be augmented intelligence. I sometimes wonder if it shouldn’t be just called fast computing. Computers have been pretty smart for a long time and is this the inevitable outcome, is the way that they can combine so many different data sources and pull together and infer so many insights out of these interconnected pieces of information. It’s coming like a freight train, but it’s been coming I think a long time.

Yvette Bohanan:

All right. Well there you have it. Time will tell. Frank, Russ, thank you so much for joining me on this conversation, the opening gambit, the top of the first inning, one out. I like that and I hope that we can get together a year from now and just see how far we’ve come, how much progress has been made, how many interesting conversations we get to have as a result of this episode airing, and I look forward to it.

Frank Young:

Excellent. Thank you. Thank you, Russ.

Russ Jones:

Thank you, Frank. Thank, Yvette.

Yvette Bohanan:

And to all of you listening, thanks for joining us. Until next time, keep up the good work. Bye for now. If you enjoyed Payments on Fire, someone else might too. So please feel free to share this podcast on your favorite social media outlet. Payments on Fire is a production of Glenbrook Partners. Glenbrook is a leading global consulting and education firm to the payments industry. Learn more and connect with us by visiting our website at glenbrook.com. All opinions expressed on our podcast are those of our hosts and guests. While companies featured or mentioned on our show may be clients of Glenbrook, Glenbrook receives no compensation for podcasts. No mention of any company or specific offering should be construed as an endorsement of that company’s products or services.