Alex Habet welcomes Dallas Wells back to the podcast to catch up on the latest hot topics in bank technology, including AI, fraud, and payments.
Listen
Watch
Subscribe
Transcript
Alex Habet
Hi, and welcome to The Purposeful Banker, the leading commercial banking podcast brought to you by Q2, where we discuss the big topics on the minds of today's best bankers. I'm Alex Habet.
Today I have a long overdue catch-up conversation with Dallas Wells. I really enjoy these conversations a lot. Typically, we start around some narrow framework and it expands into all kinds of interesting arenas and directions. We hit on things like transformation, latest frontiers, and of course AI. Not too much, but a little bit of it. But I'm also excited to share that Dallas will be making more appearances on the show going forward. As many will recall, longtime listeners will know that Dallas was a host back in the day of The Purposeful Banker. He's been busy running the product organization here at Q2 for the last couple of years, so we tried to leave him alone as much as possible to give him space to do that work. But lucky for us, a couple of years later, he's making a little bit more time for us here at The Purposeful Banker again. So stay tuned for more, but for now, hope you sit back, relax, and enjoy my conversation with Dallas Wells.
In the last year on this show we’ve rehashed a lot of what's happened in the industry, but we haven't actually spent nearly as much time outside of the AI stuff, which has captured a lot of attention. We haven't really spent too much time focusing on the non-AI developments, specifically in the technology space and things like that. So I'd be interested in your perspective. Where have you spent probably in the last year the most time focused on or trying to solve? Outside of the AI problem, which we probably need to spend a dedicated section on that.
Dallas Wells
Yeah. We'll come back to that, but outside of that ... And sometimes it's a little hard to untangle those, but I think I get what you're after. So one is, well, I would say maybe thematically it is that in many cases the technology is finally catching up to the vision that's been cast out there for a while. I'd say two big examples of that: One is the concept of personalization or being able to shape a digital experience for either an actual individual or for at least a cohort that shares enough traits to behave pretty similarly. Everyone has fundamentally agreed that that's a good thing and that there's a positive ROI in that. So the users of the technology get great benefit from it, and you get better as a provider of those things as a financial institution or a streaming application, whatever it is, there's real commercial outcomes for the providers of those things as well.
Nobody's argued that. That's really hard to do at scale. And so I think we're finally getting to the place where the infrastructure and the use of the data is in a place where you can make that real in production at scale, across millions of users. And I think the Amazons and the Netflixes of the world, they made it no longer an optional thing. That's become the expectation is that you know a lot about me. Why can't you just skip through some of this stuff that I shouldn't have to do or that I already told you before, whatever the case may be. So that's one big bucket is personalization.
The other one, maybe a little more specific to the financial and banking universe is that commercial is finally moving from analog to digital, and it's a slow, painful process because of all the complexity and because of all the custom stuff that's out there. But large complicated businesses are finally moving much more of their daily interactions with their financial institution to a truly digital self-service channel, and those that are more connected and have almost a little bit more of a consumer feel to them in a way that starts to drive some real efficiency.
But I'll give you an example. If you're a large commercial customer and you want to change financial institutions, even if that bank or credit union that you're moving to has an online application and they're like, "Hey, open your account and transfer all your stuff over." That's actually been this really complicated Rube Goldberg machine forever where, yeah, we'll take the information, which is basically fill out this PDF, and then the reality behind the scenes is that Susie in Operations is going to print out that form and then basically walk across the factory floor plugging all the wires for all the services you're going to need, and it may take six months and a whole bunch of trial and error to get all that stuff actually functioning, moving. And so everybody looked at that process, they're like, "Man, heck with that. I don't have room for that. There's not enough upside for me to go through that pain."
And we are in bits and pieces now actually moving beyond that where we can actually move that whole process to the digital world, including a lot of the supporting back-office stuff that the financial institutions have to do. What's been really interesting about it is what that has required is now that the technology exists, as banks are trying to roll this out, they're like, "Well, hey, we've always had Susie in Operations put this information in that system over there as step seven of the process. And the way you have this set up, now it's going into a different system and then an API is moving it down the tracks. We'd still really like for it to be put in that system at step number seven." So everybody wants to basically put a digital wrapper on their existing analog. I'll call it process debt. It's like the human being version of tech debt.
And so that, I think is the next hurdle to clear. And some institutions have gotten really good at it. They're like, "Hey, we are actually doing a real …" (to use the buzzword for it) … "We're doing a real digital transformation here. We're lighting all the processes on fire and burning them to the ground. We're going to start over and build a process around the new digital tools that exist instead of vice versa." But that's still not the majority. The majority of FIs are still like, "Nope. We've always done it this way. We want the digital tools, we want some efficiency out of it, but it has to match our workflows." So that's the friction point. And you feel it in the delivery of these projects. So you see a demo that's amazing because the technology is caught up and then the implementation of it gets painful as we run head on into those things.
And one more that I don't have a lot of maybe special wisdom to share other than just, like, we've spent a lot of time on it, as has everyone, just fraud. Holy cow. That surface area has gotten wide and deep. And so, again, as the technology moves faster, as there's more payment rails as real-time payments and more of these P2P networks pop up, everybody gets really excited to add those because their customers are asking for it. And it's just another fraud channel every time we open up one of those. And so there's this constant trade-off between fighting the fraud versus adding too much friction or restrictiveness to the process. That's the line every single day we’re wrestling with as more and more of our customers, including some business customers at this point, have exposure to fintechs that they're like, "Look, we got VC money to burn on the fraud side of it. We're removing all the friction." That's one set of now expectations that you have to measure against versus the reality of a regulated financial entity having to say, “But we still have to protect the dollars and cents on your behalf and on ours, so that's tricky.” But that's probably three big areas that over the last year, me and the teams here have spent a whole bunch of time on.
Alex Habet
So on the fraud thing, what is a typical acceptable level of fraud before friction? What would be a ballpark?
Dallas Wells
So here's basically what's the proper appetite for that? And it's like ... Again, you and I both came up in banks earlier in our career and it's like all loan losses are bad and all fraud is bad, and you have a bad taste in your mouth about both of those things. But if you just try to back up from that a little bit and just look at P&L just at what is driving the revenues and the expenses through the financial statements. If you have zero loan losses or you have zero fraud losses, you're doing it wrong and you are leaving substantial revenues on the table. So there's not an across-the-board answer for that other than it's not zero. And that for some ... We talked to a lot of executives who are like, "We're just getting killed by this fraud." "OK. Well, how much fraud have you had over the last six months?" "It's like 8,000 bucks." "OK. You're not getting killed by it. We spent more on this phone call to talk about it. Just eat it and move on."
There's others that are like, they have to whisper the number because they're like, "You wouldn't believe the amount of Zelle fraud that we have. How long before we just close that down entirely because it's not worth it to us?" So it's a moving target. It has more emotion maybe than it should instead of analytics. But I think that's where the institutions that we see handling it best have just ... Just like they manage their loan loss reserve and their tolerance for like, yeah, there's a certain number of basis points that's going to be about right. And we will fiddle with that number constantly and feel it out based on the cycle and what our current strategy is and all those things. But there's a range that we feel like is our happy place and we want to manage to that range rather than just we want it to be zero and anything above that causes some discomfort. And we basically just go until the discomfort gets to where there's heads rolling about it from the board level down or whatever it may be. It's basically like you have to decide upfront: Yeah, this is going to be where we want to manage it to and then be willing to adjust as the dynamics around you change as well. But it's just part of the business. It is a cost of doing this business.
Alex Habet
And when there's a new payment rail, for example, because that's happening here, I suspect the fraud tools are not quite mature in new rails as they are in older rails, right?
Dallas Wells
Yeah. That's right. And there's two aspects to that. One is that the trajectory is faster. And again, real-time payments. It's in the name is that we want this transaction to happen essentially instantaneously. So you see the benefit, you see the upside, you also see that, holy cow, when that money's gone, it's gone. That's why we have all the controls around a wire transfer that we do. Because we know from the scar tissue over the last several decades at this point, once a wire goes out the door and actually gets processed, if it's a fraudster on the other end of it, you will never get it back. They're really good at that part of it, it lands at the destination and it scatters to the winds. And so RTP will be another version of that with the difference being that we have not yet had the decades of scar tissue to build up what those controls should look like.
So we can take some of the things that we know will be the kinds of controls that we need, and we can put those in place from the outside. But we also know that as we start to get to some real critical mass with a new payment rail, that somebody's going to come up with a whole new creative way to steal money from us that we haven't seen before. So the pattern to this point has been that for each one of those new column payment rails or channels, that we come up with a very specific set of solutions to manage that particular channel. So if you're using the Visa and Mastercard card networks, there's a way to dispute those transactions. There's regulations around it and there's software solutions you can buy to manage that. We sell one of those. CentrixDTS does a great job of that.
Also, there's positive pay for checks and for ACHs. And there's new solutions that we've seen pop up around some of the P2P. Some of them are like, "We fight Zelle fraud. That's it. That's all we do." Well, as a financial institution, it's like, do we really want to manage fraud that way of every time there's a new place for people to exchange money ... like all right, I can send somebody a text message and iMessage now and I can start to exchange money that way. Does that need its own freestanding solution and therefore another line item in my bill and another person to manage it and all those things? So I think the place that we're heading to is to try to think of those less as freestanding solutions to manage each channel and instead the channels are going to ... If anything, the pace will pick up. And so we just need signals around each one of those and we need a central place to think about managing fraud that happens in our customer's accounts and a place to work cases and adjudicate those things as they come in. And instead we just pump in the new signals.
So RTP will have a new set of signals that we haven't thought we needed yet. Some vendor out there will all of a sudden supply it and it's like, here's your API surface to send us that thing so that we can now see that signal and we can make those decisions in the framework we already have instead of having to spin up a whole new one where now somebody's managing fraud and they're having to swivel their chair across 15 different applications. That's got to come through one window. So I think that's an interesting thing of, we'll see who the winners are in that. As that comes to reality. I think Q2 is making a bet on a horse or two there as we speak, but I think it'll be interesting to see how that shapes up. There's so many dollars at stake that we will come to a better place, but it's been the Wild West in terms of fraud for the last five years certainly.
Alex Habet
And it's an interesting behind the scenes look. I certainly appreciate that. On the first part, you talked about personalized I guess guidance or insights that you get. We've heard that that would be ... You often hear in the fintech echo, right, it's all about personalized experience and things like that. There's nothing new with that. But walk me through how you all decide what you're going to personalize next. How do you pick the angles of personalization at the scale of, I guess, Q2's product set to provide to customers?
Dallas Wells
Yeah. So we say personalization and even for your Netflix “things we think you'll enjoy” list, it says it's for Alex Habet, it's for people like Alex Habet. So that's the first thing is to start to figure out what are the real cohorts? And so what we've spent a fair amount of dollars and time on over the last couple of years is really getting much better telemetry on what's the actual behavior of users inside of these systems. So in the olden days, it was good enough to know how many times did they log in and how long were they in there? And that was good enough. Well, now again, we were able to do in a much more sophisticated way, when and where and how did they log in and then what was the actual flow through all the steps that they're trying to do? So that's what we're actually trying to capture is what's the different sets of behavior in there? So that's step one.
Alex Habet
Is it down to the mouse movement?
Dallas Wells
It's gotten down to where are the eyeballs. Where's the heat map on the screen of what gets noticed? You mentioned personalization. Everybody's been bragging about that for a while. What they really meant was like, "Hey, we know who to spam with email, and we know what banner ads to show them based on their behavior to this point." Now it's moving towards ... Where we're headed with it is the entire layout of the screen will be different. So we need to know for this type of customer, what should be in the top right of the screen or what should be at the top above the fold, so to speak, on the mobile app so that they don't have to scroll to it. It's like the most valuable real estate, what needs to be in that spot for these different cohorts? So it is understanding that behavior. And again, that's movement through there. It's where the focus areas are. It's how transfers between accounts are made, how payments are made, just all those behavior characteristics.
The second part of that, the one that is a little more complicated because it's so different for each institution ... We can't just do these broad across our tens of millions of users research on this, is what's actually the ideal outcome that you're after. We see what the behavior patterns are, but how do we nudge that toward the behavior that either reduces some risk, reduces some fraud, or moves that user toward completing something that we want them to do? Sometimes it's things that they started setting up a new account and got derailed somewhere. How do we get them back on track? Or they started a loan application and abandoned it or we turned them down, but there's other options for them. So we want to lead them to the right outcomes and we want to lead them to profitable outcomes for the institution. So we have to understand what those are. What's the actual outcome that you want? It's more complicated than it seems like it would be.
Do you want people to write a check for their electric bill or to pay it through the bill pay app or to put it on their debit card? Those are very different costs and revenue profiles for you for the exact same transaction to happen. And so maybe the user's indifferent, as long as the bill gets paid, who cares? But you as a financial institution might care deeply about which one of those lands out. So again, back in the early days, it was like how do we get people to do either more signature-based transactions with their debit card or more PIN-based ones depending on size of the transaction of where they're shopping and all that stuff? This stuff gets wildly complicated when you do it across, again, a customer base and all the products that you're dealing with. So that's where I think you can get lost a little bit in the noise. And so you have to start with which things really move the needle and settle on ... Not like, well, we've got 200,000 customers, we need 200,000 different experiences how to precisely optimize for each person.
No, here's about the four or five different cohorts that have very different behaviors or different approaches, and then here's the golden path for each one of those, and now we can optimize times four or five. Here's the nudges that we're making. We can measure if it works or not. You keep it to a manageable scale as you get used to it and just getting used to working this way. And what I mean by that is if you've got, all of a sudden within your institution, multiple different experiences that you're managing and you want to like, well, we'd like to test, can we get more people to use their debit cards if we change this arrangement or we send them this email or we showing this tutorial? There's lots of tools at your disposal and you need to do a little trial and error.
So first of all, you have to turn on that functionality. How do we test that internally? How do we go from a test environment to a production? Those are things that are rapidly changing. It's actually pretty easy to test in a forward environment and then just promote that to a production environment. But everybody's like, "Wait. We didn't get to do all the manual hands-on testing, and we didn't take it down overnight and rewire everything and test it again for a minute and then turn it back on. It's live on the fly. We did it at 2:38 p.m. in the afternoon during normal traffic times." That stuff, again, it's back to that process debt thing. It takes some adjustment to get used to managing that.
So as we're starting to roll this stuff out, there's lots of new technology, there's lots of new processes, and there's always a lot of enthusiasm around this of all the possibilities. And it's like, yes, we're excited too, but let's start with a handful of use cases that we can get our arms around. We can tiptoe into this approach, trip over the things that we're inevitably going to trip over in a small way instead of in a massive way. And so that's the rough steps is understanding it, understanding which way you want them to go, and then figuring out which nudges work. It's a simple thing to whiteboard out, but man, in practice, that's it. That's working with your customers now in a digital age. It used to be you would train your tellers to greet people a certain way and get faster at processing certain kinds of transactions. It's the same concept. It just now happens in your digital channels instead or in addition to in reality. And so we've got to constantly be iterating and smoothing rough edges on that.
Alex Habet
When I asked you the question in my mind, I was picturing banner ads because that is my current experience across-
Dallas Wells
That's how most places do it. Yeah.
Alex Habet
What you described sounds like magic. It really sounds incredible. How far away are we from truly being blown away when you log into your online banking and it's like this thing knows me?
Dallas Wells
Yeah. I think we're right there. So we have our personalization framework in production. There's more than a million real end users, maybe 2 million at this point, using it out in the wild. And the progression we have to go through is we start with the consumer accounts, somebody who has maybe a checking account on a CD and one loan. It's a fairly manageable relationship. And then we have to step into the, well, this person has personal accounts and they have access to their kids' accounts and they’re running a small business, and that business has 17 accounts with different ownership and different users. It spider webs in complexity there. So for the simpler accounts we're there. It exists today, we can get our arms around that cohort of what it looks like and we can give them an experience that feels very, very personalized. And then we just have to work our way up that complexity ladder.
So what we're working on rolling out next is that baby step into the small business owner who has some personal and some business stuff. We've got to figure out how to mix and match those in that experience. And then you've got people who, as we start to measure, well, which cohort are they in? What's their behavior look like? Well, it depends. Did they log in to check their personal checking account and see if they have enough to book the vacation or are they processing payroll for their business on Friday? The job to be done there can change and so what we have to figure out is how do we identify which job they're after and which experience to serve them? So those are the interesting challenges that we're walking through that we can see a pass to them, but then you go from sketching it out to doing it at scale. So we'll have some of those this year doing some of that small business stuff, which is pretty exciting.
Alex Habet
I also wonder though, as scale continues to evolve, does the compute needs to doing this widely, does that really become a bottleneck? I would imagine going through transactions and behaviors for millions and millions of customers and that can evolve and businesses can be included in that, there's got to be a lot of computing power to analyze this stuff, right?
Dallas Wells
Yeah. It is. Not to get too much into the weeds with it, but if you think about what's actually happening, you've got a user viewing something on a screen, you've got all the digital banking provider technology running its massive processes. And what we're actually doing is we're calling back to other systems inside the institution to be like, what is the current balance? What were the last five transactions? When's that next loan payment due? And then we're maybe calling out to a separate card processor to see what's the status of when you swiped at the gas station 10 minutes ago? So there's all these calls going out. So that's what the rollout looks like at this point is you build it all, the code all works, it works great, it's performative, it's stable, and then you put it into production for an institution and they test it. And what you'll find is like, all right, most of it's working as expected, but for this core provider, when they get to 12 accounts instead of 10, there's too much latency. That compute or the calls are taking too much time. So then we have to go back and rebuild how we're doing some of those calls and that stuff.
So those are the things that we're working through. And basically that universe starts big of like, oh my gosh, there's all these performance things and we just shrink them down. We Pareto them out and deal with the biggest culprits first. So with our early adopters, we're down to the corner cases, but that's a challenge that's fundamental to our business. So the framework's a little different in that we have an additional layer to go through of what audience is this particular user who's logging in. What audience are they in, and then what experience has the financial institution configured for that particular audience?
So there's one extra call in there now it's a complicated one, but it's one of many dozens or hundreds that we're making at any given time. So it does add a little layer of complexity, but most of that complexity already existed, and it's something we've been dealing with for 20 years. And not just Q2. The whole industry has been dealing with that for 20 years. So now I'm not smart enough to figure out how they solved that, but we've got teams that have been doing that for a long time and are pretty darn good at it.
Alex Habet
Absolutely. Hey, what do you think of all these copilots out there?
Dallas Wells
I think it's a step change on the order of mobile or cloud. I really believe that. We've had the benefit of seeing from the PrecisionLender days, we rolled out a copilot in 2017, so we've been able to see the power of that. Now, not the Generative AI version, but more of the deterministic rules-based version of a copilot of if you see the user doing this, guide them to do this thing, or here's the information they're going to need. And to walk alongside a user, smoothing out the process for them or helping them connect disparate systems and parts of the workflow. Grab this information from Salesforce and drop it in here, do what you need to do with it, and then here's the thing that's going to need to be passed along to the loan origination system. And the copilot can ... what was really useful about it is it puts a boundary or a handle around this set of functionality. It helps a human understand, oh, I need the copilot to go grab this thing from me and do that. Or while I'm doing this over here, maybe the copilot can sort through this data and pull out these things for me across a database of 2 million things that a human can't do, but the copilot can go grab it.
I think the technology feels a little bit like magic sometimes, but to me, the framework, it's a third party entering the process now. You have the human, you have the core software application that they're actually working in, and now you have this third thing, this third entity dropped in there that's like, "Hey, I can see what you're doing and I can see what the application's doing and I can help you to coordinate and do the job that you're trying to do." And so it's super powerful. And then when you add the Generative AI component to it it just explodes the possibility out there. So I'll tell one thing that's been interesting since we started so early, we use some off-the-shelf language processing tools for the chat interface that our copilot had, and they were really clunky. They didn't speak banker-ease. They didn't understand some of the terminology that a banker would be using. So we had to basically scrap that and build our own. And so in a very narrow set of potential things to do that copilot could be really, really good. We could program it to speak the sorts of things that it needed to be able to do.
But if you got outside of the typical rails and just be like, look, nobody's ever taught me that work. I don't know what that means. And so now when you can drop in Generative AI where it's like, yeah, I can infer what I think you mean, it opens up the possibilities in a really amazing way. So we're making a big bet on it. We've seen real evidence that it moves the needle.
It absolutely can alter behavior for the humans at scale where you can have one person make a decision. Now the copilot tells all thousands of users that are using that thing how to make that slight nuanced adjustment or how to use that data point in a certain way. It absolutely matters. And so the process we're walking through now is we have the framework, we have the platform, this copilot platform that's very mature at this point. It's been used to price at this point trillions of dollars worth of loans and deposits and evaluate profitability, and now it's doing fraud management stuff inside of other solutions.
Alex Habet
Hallucination free, I might add.
Dallas Wells
Yes. Hallucination free. Again, the deterministic approach, we know what the outcome's going to be because we've hard coded it that way. And so the question is, for many of the jobs and for many of the use cases, deterministic is the right answer. There are some jobs where that probabilistic more open-ended Generative AI thing is the answer. And it's not just the chat interfaces. That is an aspect of it. But it's actually more of the automation. Read everything that's on this form that you just uploaded and see if you can figure out what the free cash flow looks like and then load that over here. And so the deterministic thing couldn't, without a doubt, identify those, but Generative AI can be like, "Hey, I'm 99% sure that this is what you need and where you need it." So we’ve got to figure out what those use cases are of which tool to use, in which case. What we see from a lot of vendors is it's like, "Man, ChatGPT is the greatest thing since sliced bread so we're using it for all the jobs." That's not a realistic approach to use in the banking industry. Instead, it's got to be like, no, there's a lot of times where the model underneath has to be auditable and predictable, and we have to know exactly what's going to come out and we can't go outside of the bounds of what's safety or, frankly, legal.
And then there's other places where it's like, no, that either is an appropriate place where there's a human still in the loop that can get help from the Generative AI model output and just use it as another signal. Well, the model thinks this. Does that make sense? And then I confirm and go ahead. So mixing and matching those tools in a thoughtful way and making sure that our platform can use both side by side depending on where it's appropriate and clearly identify which is which and document it and all the things that you're going to have to do to keep your examiners happy. So we think it's really exciting. And the nature of the technology is that you can see some of the cool demos and things that are out there and then pretty quickly roll them out in a real way. It's like there's a new version of ChatGPT available. We test it. We connect it to whatever data model we're comfortable with exposing it to, and then off you go. So I think it'll be a fast innovation cycle that will blow people's hair back a little bit.
Alex Habet
Yeah. I didn't really contemplate the notion of some generative models in certain workflows and then purposefully excluding generative models in other contexts because of the risks that they bring and having to be a hundred percent correct almost excludes them in certain cases. But wouldn't that in theory become ... That's like a whole new area to manage for banks at this point. They don't have this discipline today as far as I know. They're good at managing models, but they're not good at managing this model yet.
Dallas Wells
Yeah. That's right. And I think that's where it's made it ... Everyone's talking about it. They're paying attention to it. They want to know, “Hey, show us what you're doing with Generative AI,” and we show them these amazing demos. And we're like, "Here's this thing, and we're actually pretty confident in it. Do you want to buy it?" Like, "Oh God, no. I just wanted to see what you were doing. I'm not ready for that yet." So I think that's what'll be interesting is it's like real-time payments. We talked about that a little earlier. The adoption of that has been pretty anemic to this point. And so it'll happen slowly and then all at once. And I think Generative AI will be the same way. Some of it is a comfort level of which approach are you comfortable with and is your board comfortable with? And can you get your arms around the risk in the way that you responsibly should?
And so I think rather than just say, here's your Generative AI, plug it in and use it for everything. The Microsoft Copilot approach of like, hey, it's in there. Once you turn it on, it's loose in the enterprise. I think for banking, it's got to be much more of a use case by use case. So in our terminology, there's the Andi Copilot and Andi has skills. And so our intention is that certain skills where it's appropriate will be ones that use Generative AI. But you opt into those and you decide, yes, that's a skill that the outcome is worth it. And I have well documented risk controls around it. I will turn that one on and use it for just that thing. But everything else stays this deterministic, a little bit more of a closed system, better controlled approach to it. And my guess is that you'll pretty soon wake up and you'll start with one, and all of a sudden you've got 101 of those Generative AI skills. But it's like making that conscious decision use case by use case. This is the right tool. We understand it, we're comfortable with it. Let's go.
Alex Habet
At first when I experienced things like Generative AI the first time, personally, I thought that it would take over everything within six months. How could it not? But it's now been a year and a half-ish. And there is adoption on the consumer side, I think at some level of scale, but in the industry it's still not there. And that's fine for all the reasons we talked about. Do you think though, in five years-
Dallas Wells
I think it's everywhere. Yeah. In five-
Alex Habet
Do you think it's everywhere in five years?
Dallas Wells
Yeah. I think it is. It's too game-changing to not be. Lots of people want to understand where we are as a tech company, where we're headed with it, and it's not so that they can buy it next quarter, it's so that they are comfortable with the fact that, yeah, we've got partners that when we're ready in two years, three years, they'll be ready also. And that's, I think, the place everybody's trying to get to. So it's going to be a really interesting ride to see how that adoption actually happens. But I think it's going to be everywhere. There's too much efficiency there to be had.
Alex Habet
All right. Well, I have to remind future Alex to pull this footage in five years and have a conversation with you.
Dallas Wells
Check back how dumb we were, how much we nailed it. Yeah.
Alex Habet
All right. Well Dallas, thank you. Thank you again for coming back on the show. Long time, but it's been great to catch up. And from what I hear, maybe you'll be coming back a little bit more regularly.
Dallas Wells
That's the plan.
Alex Habet
Yeah.
Dallas Wells
We've rearranged some duties and part of it was to make some more time for this. So look forward to being back more often.
Alex Habet
I think a lot of the old listeners of the show are going to be very pleased. We're bringing back the old crew. Like Jim Young came back.
Dallas Wells
We'll see how the data goes. Yeah. If you see me for two and then I'm gone, you'll know it didn't go well and it was worth a try.
Alex Habet
I'm sure there'll be a huge spike and you're locked in on a contract.
Dallas Wells
All right.
Alex Habet
All right, well thank you again and we'll catch up soon.
Dallas Wells
Yeah. Thanks Alex.
Alex Habet
And that's it for this week's episode of The Purposeful Banker. If you want to catch more episodes of the show, please subscribe wherever you like to listen to podcasts, including Apple, Spotify, Stitcher, and iHeartRadio. And if you prefer a video version, you can find us on YouTube, as well. As always, let us know what you think in the comments. You can also head over to q2.com to learn more about the company behind the content. Until next time, this is Alex Habet. You've been listening to The Purposeful Banker.
