Thinking About the Future: Structure Your Thinking

Announcer  00:05

Welcome to the Treasury Update Podcast presented by Strategic Treasurer. Your source for interesting treasury news, analysis, and insights in your car, at the gym, or wherever you decide to tune in.

Jonathan Jeffery  00:18

Welcome back to the Treasury Update Podcast. I’m Jonathan, media production specialist here, and I am talking with Craig Jeffery. Welcome back to the show, Craig,

Craig Jeffery  00:27

Thanks, Jon. I am really looking forward to today’s discussion.

Jonathan Jeffery  00:31

Today, Craig put together a couple of items to guide our thinking as you’re thinking about the future, looking at changes, and a couple different items before we move into the list of items you’ve got here. How do we understand changes so that we don’t dismiss something that’s useful or adopt something that is going to be archaic pretty soon.

Craig Jeffery  00:51

When there’s a hypothesis, there can be type one and type two errors, like your test or your conclusion. Can dismiss something like you said, that is valid, and you could also accept something as true that’s false. And so just thinking about how we look at changes that are taking place and evaluate them properly or calibrate them properly, is is really essential, because we oftentimes have success by thinking quickly, coming to conclusions by relating those to other actions or other events, other products, other things that may have been hype at some point in time, and then make make a quick decision that may be right or may be wrong. And so what I want to encourage people to do is think about things a little more deeply so we don’t miss some significant opportunities. Skepticism of things, of ideas and plans and new technologies is useful, but we also have to have a reasonable amount of skepticism about our skepticism. There’s a lot of hype, and sometimes the hype is real and sometimes the hype is real, and the opportunity is great, it just may be maybe on a different time horizon.

Jonathan Jeffery  02:06

Okay, and the first item you have here sounds a little bit nonsensical, but hopefully you can make it make sense. What do you mean by small is big?

Craig Jeffery  02:14

Yeah, I could have said small is big, or small is fast, or small is versatile. That’s really talking about technology. So for the non technologist, I’ll describe it this way. I think I’ve talked about this in different webinars at different times, but if you have been around for a while, some of the older technology, there would be upgrades, like every 18 months, and there were, there would be certain layers, the database layer, the user interface, layer the business process, layers that might be three tiered architecture and tier if there’s multiple layers, and the updates would be very far apart. And so now we see in the SaaS world, updates coming more quickly, every four weeks, or every six weeks, sometimes every week, changes are coming, and what I mean by small is big. When we break these items down to smaller pieces, it allows for faster development. I’ll just give an example of something like micro services. If everything is tied in together, very significantly, changing something in the system is very hard because everything’s bound together, but if we break things down into smaller and smaller pieces, you can get to the point where those pieces can even be replaced while the system’s running. You can turn off access to a certain piece, drop in a new container or micro service, or some other element, that would be an enhancement, even while the system’s running. And so this breaking things down into smaller pieces, you know, is one of the key areas why development is moving so much faster. Instead of everything being bound together, you have to unpack it all, change one piece, bind it all up together, makes it move much more rapidly. And so the newer technology leverages all of these, these tools. It allows you to make smaller changes and increments. And what this means, the end result in this is development is often five to eight times faster, you know, let’s say, using cloud native than traditional software as a service, even one generation behind technology. It’s, you know, takes quite a bit longer. And it’s, it’s multiple times, five to eight times, not, not 10 times, not 20 times, but significantly different. So small is fast, or small is big, that’s much more powerful than I think many of us would have thought. Yeah, with the small updates over time, you’re probably learning a lot more learning opportunities than gathering everything up for a big update. Yeah, and it’s in its and you can leverage and use more common tools and put them together in ways that make sense. So, and there’s been different examples you could you could talk about documented, undocumented calls and operating systems. And so if someone’s building a program and they can use a documented call, you know, for sorting or graphing. Well, all I have to do is refer to that. Now I’ve leveraged that other tool set in my development that would be the small is bigger the operating system versus the individual program. That would have been what we’ve talked about 10 or 15 years ago. And so I would just describe it this way. It allows us to rapidly pull these things together and develop far more rapidly than we could under the old technology structure.

Jonathan Jeffery  05:30

Okay, that started to make a lot more sense, and I think I’m going to say that when I’m fishing now, whenever I catch small fish, small is big. I caught nine small ones.

Craig Jeffery  05:40

That would be a, that would be a blatant misunderstanding of, small is big. I’m hungry. Hungry is full. I mean, it’s like, it’s the small is it’s a bigger deal. A little bit equivocation in terms of, like, smaller components can lead to a bigger impact because of how they’re put together. So, yeah, it’s a bit plain words with the way I say that, but yes, small is more powerful. Might be a more accurate, but small is big. It’s like, what does that mean?

Jonathan Jeffery  06:06

That’s more catchy. Okay, the second one is, SaaS is different than platform as a service. Yeah.

Craig Jeffery  06:13

So Software as a Service is this idea that I can have multiple tenants or individual tenants, but all the updates to the software are done, and then all the users of that software realize the benefits of that like their software is maintained. It’s updated, it’s regularly improved upon. And that’s a way of making the software, your software investment, your technology investment, grow in value over time, because I’m not sitting there waiting for a massive update, and then I don’t have any value for that update, making the software obsolete. So Software as a Service has been around for a while. That’s not the same thing as a platform as a service or pass so those two things together would make up what we would call cloud native platform as a service. You can think of Microsoft Azure, Amazon Web Services. They offer platforms where there’s all these tools that exist. Now, just because you say Microsoft Azure, Google Cloud, Amazon Web Services, just because you can mention the different service providers doesn’t mean that it’s being used as platform as a service. And an issue may be I just need to host something. I’m going to run it on the cloud, and I’m going to run it on Amazon Web Services or Azure. That does not necessarily mean that you’re taking advantage of the Platform as a Service technology. And so I’m speaking to the lay person here in terms of how it used. So SaaS is different from from PaaS then.  You can have software as a service running, you know, in a variety of different ways, but these two items together is the definition of cloud native. There are multiple definitions of cloud native, but it’s using the most current tools and techniques to develop it in a common way, on a highly scalable way, using the most current technology. So the platform versus I just port something over into the cloud that I ran locally. This is a different generation of that technology.

Jonathan Jeffery  08:18

Where are you seeing this debate coming up?

Craig Jeffery  08:22

Well, I don’t hear that debate coming up much in the treasury space, but the but examples of this would be if I have a platform that supports things like artificial intelligence, natural language search, different types of scalability, where it can, you know, rapidly scale up and down. That’s very different on a platform level. If a normal day I need, as a vendor, I need 100 units of compute, and then at the end of the month, I’m running hedging activity, and I need 4000 or 40,000 units of compute. I’m not just running it, and I have a reserved amount of capacity that I can use. So I have to go, I have to, you know, always have double what I need, so that when I hit a heavy period of time, it doesn’t slow down too much. If I’m using the newer tech, I can scale up the speed it goes from, you know, 100 I think was my example, of 4000 that’s needed for three hours of processing, and then it drops down so nobody’s hitting that constraint, it expands or contracts as needed. So the platform allows me massive flexibility to use new tools and to manage things like compute, storage. Another example might be, I want to use natural language search for asking my questions as opposed to having here’s my help, my help options? Well, that’s great. There are help, the help options are context sensitive. They allow me to find out what’s what’s available there, but the search parameters tend to be built into, you know, people are building that so that when the user types in something, it references, it indexes it.  But natural language search uses like when you’re going to Google or Bing, you’re typing who were the last three presidents, whatever you know, you type that, and it goes and looks and hunts either in tables, free form fields, unstructured data, and comes up. And so now it can query and interrogate all of the data in your system and come up with a response. So the language that you’re using is English in this case, so it’s natural language. I’m saying what I want. The system can search everything, understand both the context for what you’re asking as well as the output style you’re looking for, and return that to you.

Jonathan Jeffery  10:44

And interestingly enough, I don’t know if you know this, but I recorded a natural language processing Coffee Break session with Paul. I just edited it this morning.

Craig Jeffery  10:52

Well, now I know, the more you know.

Jonathan Jeffery  10:54

It links together. The next one on the list is speed. You said speed matters for use cases, and nothing will be slower.

Craig Jeffery  11:03

Yeah. So that’s pretty much the whole gist of this.  When we moved and started having these new payment rails created, right for a long time we were we had things like low value payments in the US, like ACH, high value payments like wire transfers. There weren’t a lot of new payment rails in the corporate payment space, there are multiple ones in the US now, and dozens and dozens of new payment rails across the globe. And so instant payments, immediate payments, Faster Payments, branded and unbranded names exist for faster payments. This idea of greater level of speed, speed certainly matters significantly in certain use cases. Let’s say for certain types of payments, faster is way better, but not always. You don’t necessarily want to pay someone faster, but you want a richer experience with the with your trading partner. Maybe I need more information. More information may be more important than speed. So the idea that there’s use cases where speed matters, where speed is perhaps the number one, you know, speed or speed and security, those might be the number one and number two, characteristics of what matters. But certain use cases have different drivers, speed, security, level of information, ability to change, resilience, what have you, nothing’s moving slower. We don’t see anything where things are getting slower. Not everything needs to be instant. Everything’s moving faster. The pressure for faster varies over time, and that’s going to relate to the next topic that we had jotted down, that you’re going to ask about as well. So nothing’s moving slower. There are use cases for speed, but better is an improvement upon just speed alone. Better may be describing speed, level of information, level of security, ability to flex and scale. All those things might be more appropriate descriptions of better than just speed. Not everything is about speed.

Jonathan Jeffery  13:04

From batch to streaming.

Craig Jeffery  13:06

Yeah, so this is, this is the idea of, let’s say, when we think about payment files and we think about information reporting from banks, what happened to the night? Well, there’s a lot of batch processes that occur. So there’s a nightly cycle that runs at a bank, and they they grab everything together. It says, Here are the transactions that occurred yesterday, and they send that information to you, to your company. It says, Here are all the summary transactions. Here all the detailed transactions. And so that might be provided in a batch environment. Same thing. If you’re initiating a payroll file, you might send a file to the bank that says, here’s the, you know, 455 employees send it out with this type of format, and it’s sent in a batch. All 400 of those transactions are together. So that’s what we think about when, when there’s batch, it’s very efficient. It’s done after a certain amount of work has been completed. It’s totaled, it’s sent over. The difference between batch and streaming can occur inside an organization, as well as between an organization and their fintech provider, the organization and their bank. So streaming, we would think of if you’re connecting to your bank and you want to get information anytime a transaction hits your bank account. Let’s say a wire comes in, there might be an API set up that instantly passes it through when it hits or within a matter of seconds, let’s say less than seven seconds, that information is then pushed out. That’s a that’s not truly streaming, but it’s that sort of immediate update, as opposed to, we’re going to update you once a day to we’re going to update you when something happens. When something happens, tell me what it is, versus send me a summary of everything that happened yesterday, or send me all the details from yesterday at six in the morning. And so we think about that on APIs. We would refer to as APIs streaming. I’ll say this more broadly as. So instead of passing things, whether it’s inside a company or outside, that you might have an API gateway to enable this, but this idea that every time something occurs, or for the most part, when things occur, transactions hit, they’re passed to the right system or area that needs to have, that as they occur, it moves, as opposed to this is this occurs at two in the afternoon, or we have it every two hours we send a file from one place to another to update, AP, AR, treasury, forecasting, FP&A, whatever those things are, we’re moving to an environment where things are updated more regularly. Everything’s moving faster, as we said before, but this idea and the direction is we need to know stuff in a more real time basis. We don’t need to know everything immediately, but the shifting of, hey, these things are set at batch on a monthly basis, on a daily basis, every four hours to as they occur, that type of structure will change in organizations and between organizations and their service providers, their bank, their fintech, and so that’s that’s where things are going. So APIs drive that different types of gateways, it becomes more of a streaming, as it happens, it’s flowing around the organization.

Jonathan Jeffery  16:18

I can see the benefit of that on a personal side, with personal finance apps that APIs that track everything. One of them that I use updates almost every day. Sometimes it’s every two days. The other one updates a couple times a day. And it’s so much easier to stay on top of things when you see what’s coming in.

Craig Jeffery  16:35

And that’s a good example. And so you don’t need everything. So if you have an app on your phone, you’re doing a rideshare, you’re going somewhere with the rideshare service. It’s like, okay, that’s updating you to show you where the car is, how close it is to you, yeah, and that’s pretty useful, so you don’t get impatient. It’s like, okay, here’s where it is, and it’s giving you an update how long you think it’ll take there. And so now you have information you can you know, just know it, or it may impact what you’re doing. Maybe you recognize you have to call ahead of time say it looks like I’m going to be 10 minutes late when it says you’re going to arrive there, you know, three minutes after the hour. And so you’re building a little bit of a buffer in there. So it’s useful. Same is true for Treasury. It’s like we’re looking for a payment. Do I keep going out to the bank portal to ping it? Do I have that? Send it to me as soon as it comes in either as a message, or do I have it sent to my system so it alerts me that, hey, this large wire came in at a particular time, so the needs, the needs can vary, but, but that’s where that comes from. Yeah, the next one is, data is the new currency. And that’s, uh, funny too, because why do we say data is the new currency? What’s what’s good about currency? It’s valuable. Well, cash is always king, but data, data becomes more important than transactions. I think a lot of banks are realizing this, that the data that flows through transactions and the data that happens can become more valuable, is more valuable to the sender and the recipient, but also valuable to the bank, and so there is a far more significance, there’s far more significance to data that flows with a payment or with information then, then just the fact that there was a transaction, because it impacts the updating the system from an accounting perspective. But that data can also be used for additional analysis, so more insights might be able to be achieved, insights that you may be able to define today, but as you have more data that allows you to learn and do better over time.

Jonathan Jeffery  18:34

Does data suffer from inflation?

Craig Jeffery  18:36

Well, data is definitely experiencing hyperinflation in terms of growth. It’s a the amount of data that we have essentially doubles about every year and a half or two years. I mean, just a rapid expanse of data, exhaust data, data from systems we’re exporting more, creating more, everything from data that might be on the internet of things to you know all the way to you know more richer data is flowing through systems, and it can be used for for more analysis. It can be used for artificial intelligence. Your business intelligence tools can help you do manual analysis. Your ability to leverage AI and machine learning can help you see correlations of data, cause and effect, dependent, independent variables. There’s just much that can be done with that.

Jonathan Jeffery  19:24

And how should we structure our thinking with skepticism?

Craig Jeffery  19:29

Well, you know, we we’re oftentimes really effective at making quick decisions because we associate it with prior experiences, and our prior experiences are really valuable for making good decisions. This looks unsafe. I’m not doing that that protects me, as our ability to to make those decisions quickly has been a strength, and so being skeptical of your skepticism is that it is easy to make an incorrect assumption on something, and so don’t just that could be your operating assumption is that this is no good. This is risky. We should do this, but I want to do some additional validation, depending on what the potential outcome is. So being skeptical, I’ll just give an example.  Electronic bank account management was this panecea to the issue of, we’ve got to open up accounts. KYC is a pain. There’s all this back and forth between the company and the bank. Open this account, here’s the signers, here’s the signatories, here’s the documents that go with it. This back and forth. KYC was a pain. And so there’s this development of, here are all the conversations and the individual messages that exist, and they were all defined, and that can be sent through Swift it can be sent back and forth between host to host. And was that valuable or not? Well, we can say that there was a lot of hype because most companies didn’t have their their data organized. It was in paper files, maybe multiple spreadsheets, some word documents stored in different places or stored in physical files, also on people’s desk. And so it never took off. It hasn’t taken off.  It may, at some point, but you can’t interface paper into a digital system, paper and stuff stored everywhere into a digital system or a conversation into a workflow. And so we learned from that that sometimes there’s hype around something, there’s no way to do it. In this case, largely the prerequisites of the foundational items that were necessary for that, the structural things were necessary didn’t exist, and so you couldn’t do that. You know, it’s like, I’ve got a water based system. You’ve got a electronic based system. Those things won’t interface, not well, not without a lot of shock. And so there’s not this ability to to make use of it. And so we can say, Oh, this other thing looks like something else that failed, therefore I can ignore it. And I’ll give you the other example. Think about the fake emails, the phishing emails you used to get. I am prince from, list whatever country, you know my uncle or I’m a member of the bank and somebody died. And they’ll reference some, some tragic event where their plane went down. And they’ll say, I need help getting this money out, and do you have an account strong enough for it? And the phrasing was horrible. It wasn’t, wasn’t written well, it was like someone was typing with the hammer and a cucumber just banging the keyboard. And you’re like, okay, I understood what you said. Who would fall for that? Nobody would fall for that. This is ridiculous. And then it gets to the point where some people fall for it, right? Somebody that might be susceptible to it to now, with the use of AI and these managed processes, the phishing emails and the business email compromise, the communication that comes out is extremely good. I mean, it might use the same phrases, same type of language. It’s they’ve learned a lot, and so I don’t think anybody would say that’s not a risk now it’s a huge risk because it’s so well done. So what started as horrible is now extremely convincing. And so we have to step up our game.  So we could ignore it when it was terrible. We can’t ignore it now. And so that’s an example of things change over time. So review your thinking. You know, AI, is that hype? I don’t know. How would we use it in treasury? Well, it’s good for words. Is it good for analysis? Maybe it’ll solve everything. Maybe, maybe it won’t. But this idea of thinking through, what are the implications and what will be necessary?  You know, AI, for example, when it first comes out, you’re like, I asked questions about certain things, and it was horrible. I refined my questions. I got it to be a little bit smarter, and it was way better than an intern, but I still can’t rely on it. And so if you project that to the future and say, this won’t work ever, that’s a very different thing from maybe it’s not ready to be relied on fully. You know, as an example of the computer chess games right that were built to beat Kasperov, but Deep Blue, the second time they played, beat him, they they took it away, and then they had games that you could load, every single that loaded every single chess game that we had records of, in addition to logic, and they became essentially unbeatable, or very, very close to unbeatable. And then, using AI, they built the rules into the system, and they allowed the computer to play itself time after time, millions of times in a nine hour period. And that, I think that 72 out of 100 games, or maybe as they won, 28 out of 100 and the other the rest, were draws playing against those computers that had all the games that had been played in tournaments that were captured. It won a huge percentage and tied the rest of the time. And so these things you know, AI, for example, changes dramatically. So be skeptical of your skepticism.  You can come up with conclusions, but also be mindful of those things change.  And what what made sense for how you solved the problem 10 years ago may need a new look for today.

Jonathan Jeffery  25:09

So to tie all this up, structured thinking, thinking about the future.  What key takeaways can you leave with the audience?

Craig Jeffery  25:16

I think the key thing is we have to evaluate what’s coming in front of us, and see how things are changing and why, and use that to continually refine our approach to learning, our approach to technology, our approach and thoughts on where do we find value taking those things together reflects learning from what we see with regard to changes that come our way, changes that are brought upon us, changes that are happening all around us, or changes we’re bringing to the situation.

Announcer  25:51

You’ve reached the end of another episode of the Treasury Update Podcast, be sure to follow Strategic Treasurer on LinkedIn. Just search for Strategic Treasurer. This podcast is provided for informational purposes only, and statements made by Strategic Treasurer LLC on this podcast are not intended as legal, business, consulting, or tax advice for more information, visit and bookmark StrategicTreasurer.com.