Navigating AI Hype and Reality in Banking, with Barath Narayanan

In the latest episode of Digital Banking Podcast, host Josh DeTar, VP of Sales and Marketing at Tyfone, welcomed Barath Narayanan, Global BFSI and Europe Geo Head at Persistent Systems. The episode centered around how artificial intelligence is shaping productivity, strategy, and risk management across financial services worldwide.

Barath shared how AI has moved from a side conversation to the core of every client discussion. He described a clear shift: organizations now focus less on cutting costs and more on boosting output and speed with AI tools. Barath explored how adoption varies by region, with some markets demanding higher productivity gains and others held back by legacy infrastructure and data challenges. He stressed the need for careful governance, drawing on real-world stories where moving too fast led to costly mistakes. He also highlighted the importance of having experts in the loop and building strong guardrails to protect trust and brand.

Throughout the conversation, Barath and Josh examined how evolving technology, data quality, and global perspectives impact AI’s role in digital banking. The discussion underscored that progress in AI adoption is incremental, and that staying relevant means blending bold action with thoughtful risk management. 

Transcript

[00:00:00] Barath Narayanan: Like a client of mine in the insurance sector two weeks he and his entire team spent almost a week in San Francisco and he comes back so excited about what he saw and he tells me, ETH, we are not doing anything that what I saw in San Francisco, the potential possibilities are humongous in terms of what we can adopt and the throughput of productivity can be Forex.

[00:00:28] Barath: So his point was not to reduce, his point was to increase the throughput productivity velocity with better quality and do more for his business. That is consistent, Josh, across the board where everyone wants to

[00:00:45] Barath: do more with the kind of effort you have.

[00:00:49] The Digital Banking Podcast is powered by Tyfone. Tyfone is the creator of nFinia®, a dramatically better digital banking platform for community financial institutions, as well as several platform agnostic revenue generating point solutions. Our highly configurable platform and broad ecosystem of third party partners ensure our entire suite is scalable and extensible to meet the needs of any fi.

[00:01:18] On our podcast, you'll hear host Josh DeTar discuss today's most pressing financial technology topics with seasoned industry experts from every possible discipline.

[00:01:36] Your podcast hasn't officially made it until there's ads in it, but this is one you're not gonna want to skip past. And well, if you do, feel free to hit that fast forward 15 seconds button twice, ever wonder what gives me my energy and enthusiasm during these podcasts? You know, outside of my relentless desire to learn about, connect, share, and build up community fis and their mission to support the communities they serve.

[00:02:02] It's coffee and lots of it. Now you wanna know what's better than your regular old coffee. How about donating $5 to the Children's Miracle Network Hospitals through credit unions for kids every time you purchase high quality, ethically sourced coffee that also provides living wages to coffee farmers.

[00:02:22] So if you wanna listen to this episode with epic levels of caffeine induced focus and help kids in need. Head to Java for kids.org to learn more and buy a bag or 10. Thanks.

[00:02:35] Josh DeTar: Welcome to another episode of the Digital Banking Podcast. My guest today is Barath Narayanan, the global BFSI and Europe Geo Head for Persistent Systems. Look, I got a super fun episode planned for you today. Now, while this topic is one we've covered a lot lately, my guest Barath has a very global perspective on the subject that I really think you're gonna find interesting.

[00:03:01] Josh: speaking of global perspective, how's this for Global in the last eight days? Mind you we're recording this two days before Thanksgiving, a holiday that's super important to Barath to spend at home in New York. Barath has been from New York to Tokyo, to Sydney, to Singapore, to Abu Dhabi, and is now just about to head back to New York in the morning. While this is all work related travel, I think it also speaks to a core pillar of Roth's identity. He said to me, while I'm a true Indian at heart, I like to be thought of as a global citizen. As he's evolved over the years. This global exposure taught him a few very important lessons. One is to continue to double down on being you truly unapologetically you no matter where you are.

[00:03:59] Josh: But it also taught him to learn and to grow and to take inputs from other cultures, geographies, and walks of life. And it led him to a philosophical principle that guides him as a person now, and it's the concept of being completely authentic, transparent, and accountable. He mentioned that in all his experiences in life with all the people he's met, lots of leaders talk about those characteristics, but not all of them live up to it.

[00:04:29] Josh: So his goal is to keep those as guiding principles for him every year. He's said his goal is to peel another rapper off himself to be more and more authentically himself. And as someone who described his younger self as quote average, he has seriously surpassed that in his pursuit of making a difference, really focused on the humanity of how we do work in business.

[00:04:58] Josh: Now, after 24 years at the same company, just two years ago, Thanksgiving week actually, so serendipitous of our recording today, after just thinking about taking a break from the professional world, he was convinced to jump back in and start a new project and is learning and collaborating globally on a daily basis if you ask me.

[00:05:22] Josh: That's some pretty cool stuff. So broth, thanks for joining me. Welcome to the show, sir.

[00:05:27] Barath: Thank you. Thank you so much, Josh. That's something amazing way to Really call it who I am as a

[00:05:33] Barath: person, so really appreciate what you took time to introduce to me. Thank you. Looking forward.

[00:05:38] Josh: Oh yeah, no, I, from the first time I got a chance to, to just have, a little bit of a, an introductory call with you. it was so obvious that you've just got such a unique perspective on the world, and I just, I love that about you. And it was like, every time I've talked to you, I've taken away one new interesting nugget.

[00:05:59] Josh: And what's crazy is they've all been from totally different subjects. and I actually want to start the podcast with a gift to our listeners. what's funny is that the last episode that just came out before this one does. Same thing, we end up somehow getting on the topic of parenting.

[00:06:16] Josh: and I think that's one of the things that's so cool about the podcast, medium as content, is you go all over the place. and my running joke is, I think at this point, if I had known what this podcast would become, I probably would've named it something different when we started it.

[00:06:30] Josh: 'cause what does parenting have to do with digital banking? I don't know. I'm sure we can find a connection somewhere. But anyway, I digress. you made a comment to me when we were talking about just how you think about almost your definition of success for you as a person.

[00:06:45] Josh: and you were talking about how important it was for you to make your dad proud. And, as a father now of two younger kids, like I, I think about that, but I more think about it as I want to be proud of them. And it's gonna be cool when that shifts and they want me to be proud of them.

[00:07:04] Josh: And, I'd love for you to just describe what you described for me as that kind of light bulb moment when it switched for you and you were thinking about how,somebody introducing you and your dad and that change was,

[00:07:19] Barath: Oh, that, that takes me back almost. 30 years or into when this happened. So thanks Josh. Thank you so much for up. And I'm actually, I'm more of an emotional person and when you bring this up, it calms me down to think who I am and how I grew up. So this is sometime 96 when, one of my friend's mom, very closest friend, and this is a household up, and she's another kind of a person who mentored me and brought me up during my childhood.

[00:07:49] Barath: And she made a statement, one of the evenings stating that what really matters in life is whether. You would be called as Mr. Nan's son, my dad's name is Nan. Or at some point when it turns out to be, oh, here is someone who is going his birth's father, and that is the proudest moment for both you and him.

[00:08:12] Barath: And that is something you want to keep with you Barath. And I don't know, it just stayed with me. Of course. My dad is my role model and if the hardworking and commitment to whatever I do comes with me today is from my dad, whom I've looked up and grew up throughout my, childhood and even today, I look up to him, his discipline when I'm nowhere close to his discipline,right?

[00:08:37] Barath: Your father is mostly your road model always. And that this statement from somebody who also brought me up as her own third kid made a big difference to me. today when I see my daughter and somebody, when they say, oh, here is Mata's father, my daughter's name is mea. And somebody, they say, oh, this is Me's father who's going in there.

[00:09:01] Barath: I can relate

[00:09:02] Barath: back to that and I really feel proud about it.

[00:09:06] Josh: That's cool. I know. Yeah. now

[00:09:08] Josh: I'm excited for the day where, somebody's oh, you are Jack's, Dan. wow, 'cause then you're like, yeah, then your child is the one on the pedestal and you've taken the step down. I don't know that was just. Yeah, I told you, I was like, I'm taking that away from our podcast.

[00:09:25] Josh: I'm probably gonna learn a bunch of other, like cool work related stuff. but that was a really solid takeaway that I felt like, is gonna totally change how I think about my relationship with my kids as they grow. That was really cool. So thank you. what's funny is that also kinda leads into, talking about the discipline and man, thank you so much for, somehow managing to squeeze in this episode with me.

[00:09:48] Josh: With all you've got going on, going into the holiday, loyal listeners of our podcast and followers of my LinkedIn, I get all sorts of comments all the time from people about how impressive they think my work travel is and how much I do. and then someone like you comes along and I'm like, yeah, trust me, mine's not that bad.

[00:10:08] Josh: That's, that's a whirlwind tour, sir.

[00:10:11] Barath: it is. And many of my own friends who are very close to me why do you really have to travel? Is it really you enjoy traveling? Or is it like you can't be in one place? In fact, a client made a statement yesterday. Would you get uncomfortable if you're not traveling at all? And hence start traveling, various element comes in, right?

[00:10:32] Barath: But to me, it is truly I enjoy being in front of my customers genuinely, right? And when I'm, even though I'm traveling so much, there has not been a day where I wake up and I feel uncomfortable, oh shoot, what am I going to do today? Because whatever you're doing, you're so passionate about that. And to me, travel is just an enabler for you to go and meet my clients.

[00:10:57] Barath: So for my clients conversation and meeting, I would go

[00:11:01] Barath: anywhere and that's what keeps me going,in my journey so far.

[00:11:05] Josh: I tell you what I really, I feel you on a very

[00:11:07] Josh: deep level on that statement.because I feel like I've had a very similar realization to that, even just not that terribly long ago. and it came from, I was talking to one of our customers when I was traveling and it was one of those just back to back.

[00:11:22] Josh: 'cause I hadn't been home, I hadn't seen the kids in a while. and it was hard. And, my customer asked, they were like, do you enjoy the travel? I see all the travel that you do. and and they asked me like, if you could give up one part of your job, would you give up the travel? And it made me think, and my answer actually became more important for me than I became to answering the question. And I said, actually, the very first thing that I would give up in my role if I could, would be the travel. I would give that up in a heartbeat because I would prefer to be home with my wife and my kids and here in my environment.

[00:12:00] Josh: I said, but on the flip side, the last thing I'll ever give up in my role is the travel. And it's because of what accompanies it is even in this crazy digital world, which is a lot of what we're gonna talk about today. Like even in this crazy digital world where we've got all these productivity tools, all these collaboration tools, all these connectivity tools, I would still argue there is no replacement for sitting at the table with someone.

[00:12:28] Barath: Absolutely.

[00:12:29] Josh: And so I just, I can't give it up. I have to go and sit with people. Because the types of conversations you have are very different.

[00:12:36] Barath: Even this own conversation, Josh, if it was in person, both of us sitting across the table with a glass of whatever it is, water or wine or beer or whatever it is, it makes a huge difference, right?

[00:12:49] Josh: It does.

[00:12:50] Barath: My daughter asked early, at last asking me with Zoom calls coming in, do you really need to travel?

[00:12:58] Barath: And when I said that with an example, when I told her that when you go and meet a customer, X, Y, Z, and the kind of conversation we have sitting across the table or even doing whiteboarding or different things, there is no replacement to it. And that is where you truly enjoy that. Now that is definitely a balance because when I travel, I do miss taking time out to spend with my own team members and various other things.

[00:13:25] Barath: But then I need to plan that in a different way. But for me, what gives me a kick is being in front customers

[00:13:32] Barath: and being able to solve their challenges. Gives me

[00:13:35] Barath: the kick. Yeah.

[00:13:37] Josh: Yeah. I'm with you. I know it's, It's just a different type of conversation and there's absolutely the types of conversations that can totally happen over Zoom, and should, right? and don't need to result in a high cost of doing business and expensive flights and hotels and all of these things.

[00:13:52] Josh: But at the same time, there are certain conversations and certain collaborative elements of doing business that you can't put a price on putting people in a room together,

[00:14:03] Barath: Absolutely

[00:14:04] Josh:

[00:14:04] Josh: And so I think it's cool. We live in a very cool time in human history where we can do both. But even then, broth and I were just talking about, 'cause he did the, New York to eventually Sydney.

[00:14:17] Josh: We were just talking about their, I think, what was it, late next year, they're launching the new longest commercial flight in the world, and it's 22 hours New York to Sydney. it's incredible that we live in a time where you can do that. So we can both just pick up a phone and do a FaceTime call to somebody and look 'em in the eyes and talk to 'em, which is go back not even that long ago.

[00:14:40] Josh: And that would be mind boggling to somebody. And then at the same time, we could actually put ourselves on an airplane and fly from New York to Sydney. Like it's just wild to live in that time where we can connect with people that we otherwise never would have and experience things we never would

[00:14:56] Barath: true. And all that I can say, Josh, is it's only going to keep improving York or Sydney. Timeline to travel is good to only keep producing with what we see with varying leaders doing in the industry, right as they innovate. that is going to shorten more and more, as we see the going to increase, far more, than what it is today.

[00:15:18] Josh: I'm all game for teleportation.

[00:15:20] Barath: And I was going there, but I thought.

[00:15:22] Josh: man, I tell

[00:15:23] Josh: you what, if I could just teleport to the east coast instead of flying for six and a half hours, that would be just awesome. So speaking of the changing evolution of just how we're doing business and communication and collaboration, let's rip the bandaid off and tell people.

[00:15:42] Josh: One of the topics that we plan to talk about is ai. I know, shocker, we've been talking a lot about it, but like I said, here's the thing. there's so many different facets to how it's impacting, daily lives to businesses and, I think there's so many different unique perspectives to it.

[00:15:59] Josh: and one of the reasons why I really appreciate your perspective broth is that it is, it's a very global perspective.

[00:16:06] Josh: And you're for the most part focused on the financial services sector of the world, but you're looking at other industries, you're looking at countries with totally different banking laws, regulations, ways of doing things, economies, the whole nine yards.

[00:16:23] Josh: So I think it gives a pretty cool look into kind of the wide range of how this is impacting things. I don't know, maybe just at a very high level, give me your thoughts on, what have you seen and experienced over, the last maybe 6, 12, 18 months, and what's changed and what's percolated up in, in your eyes?

[00:16:46] Barath: it is interesting. We are starting with that question, Josh, right? last maybe even 18 months or 12 to 14 months, have significantly. It changed from a point when we were having four or five conversations out of 10, which were around ai because as a firm, we are ai first couple of years ago, not recent phenomenon, right?

[00:17:11] Barath: So we did my con, my delivery counterpart, j deep doke and myself, started lot of on AI with our customers. At least five out of 10 conversations where ai, this is, mid of 20, 24,

[00:17:24] Josh: Okay.

[00:17:24] Barath: right? Today, as you with ai I can totally relate to because so much of AI conversations going on.

[00:17:32] Barath: But literally, Josh, every conversation with my customer, 10 out of 10 today is ai, 10 out of 10. There is no discussion of solving a problem without having an AI element Now. There is also a bit of a question coming into play, Josh, that a client of mine said two days ago. Are we over pivoting on AI for problems that can, that need not really need an artificial intelligence or a gen ai?

[00:18:03] Barath: It could be a simple automation, simple digitization. It doesn't really need a continuous learning at all. I agree with the client, so we are moving too much around that. But every conversation has an AI element. Now, what I saw in last, say six months to my own, my own, kind of, exposure by different regions or no how companies are adopting AI in Sydney versus Tokyo versus Singapore versus in Middle East.

[00:18:43] Barath: Or European countries or us,like by default we use AI first as a culture, as an approach into engineering work that we do, So we have a broad framework where anywhere within 25 to 30% ofJosh, we are comfortable to have a conversation with our customers. When I take this framework to US clients, they struggle a bit saying, oh, we are between 10 to 15 to 20% Or there is a legacy estate. I'm not sure I can fully relate to 30%, but are you sure that is a conversation in US banks, right? You go to Europe, it has started not entirely into production, it depends on the CIOs who are AI first, but still anywhere within 15 to five to 15%. You come to Middle East, Josh? To my own surprise, a client of mine said, Barath, what are we talking about?

[00:19:45] Barath: 30% if not minimum, 50% of productivity conversations. It's a non-starter for me.

[00:19:53] Josh: Wow.

[00:19:54] Barath: what do you mean? He said, come sit with me. Understand what we have done so far, and then if you're bringing in your own thinking process of various tools, ideas, solutions, it should be minimum where I'm today, which is around 50% I initially didn't believe.

[00:20:13] Barath: Then we spent couple of days of workshop with the customer, their team members, and they showed some demos. They showed their tools and data space. Actually, this specific client is far ahead.

[00:20:25] Barath: Then I go to Singapore and I meet a bank and this specific bank in Singapore, they were migrating around 150 roles in the credit risk space.

[00:20:36] Barath: Josh. To one of the India Center. I don't know how to say the city. Then everybody quickly guesses who the client could be, right?But they were migrating 150 rules. They decided to apply Agent AI approach to this particular process. You won't believe they told their India teams to pause the hiring.

[00:20:55] Barath: And what they needed to really scale up was only 40 member, team 42 to be precise against 150 member team. They were originally planning.

[00:21:04] Barath: And then you go to Japan. It totally varies. Still, lot of paper related, non-digital approach to work.

[00:21:13] Josh: Really?

[00:21:14] Barath: Every country, every, industry segment varies journey.

[00:21:19] Barath: if you ask.

[00:21:19] Josh: have you ever seen,it was an old like infomercial. A commercial that used to run in, on TV here in the States. And it's, I think since then become a viral meme that gets used of the tape that will stick to anything including like wet, surfaces. And the, the commercial was this like giant cylinder, clear cylinder of water with a hole and water spouting out.

[00:21:45] Josh: And the guy has the black tape on his hand and just slaps it up against it and it stops the water flowing, And now it's turned into these viral memes. And the one that I've seen a ton of recently is,a company is the water tank

[00:21:59] Josh: and they're like, all of our problems are the water pouring out.

[00:22:02] Josh: And then the guy with the tape is the CEO and then the tape is ai. And he is just slap AI on it. Just slap AI on it.

[00:22:10] Josh: and I feel like there is some of that, right? there, there is a lot of this, Oh, let's just throw AI at everything. And it's it's not really a silver bullet for everything, do we?

[00:22:20] Josh: Do we really need it here? But at the same time, I would agree. I don't have nearly the global experience that you do, but I would say even just talking to, folks in my circle, in the US financial services market, I see a crazy varying level of everything from, I've heard about it. I see some vendors with AI in the name of their booth at a conference to, people who are actually legitimately able to measure productivity gains in using some sort of artificial intelligence in their business practice.

[00:22:57] Josh: But at the same time then, one of the things I'd love to get your perspective on is I had a guest recently on the podcast a couple of episodes ago,VIT, and he is. the, chief technology, chief Experience Officer for a crypto payments company out of Dubai.

[00:23:19] Josh: He's actually a Russian guy who's been in Dubai for a long time, and he and I were talking about how his organization specifically has been using ai and he said something to me that stuck with me ever since. and I'd love to get your perspective on how you're seeing this being thought about across the regions that you're talking to. So he said, they had a conscious conversation at his company where their executives all got together and they said, how are we going to use ai? And not even just what use cases are we going to use it in?

[00:23:50] Josh: are we gonna use it to do coding? Are we going to use it just to take meeting minutes? Like, how are we gonna use it? He said, even more. We thought about it at the. High philosophical level of how is this going to change how we run our company? And he said, essentially we boiled it down to we have two paths available to us.

[00:24:07] Josh: We can use these tools to be more efficient. The paths are, how do we take that efficiency? And he said, one path is to say that we could develop a feature in six months with 10 people, but if we add AI coding to that, we can develop it in six months with two people. And he said, so now our cost goes down, right?

[00:24:35] Josh: He said, the other path is we could keep those six people, or even hire more, and then give 'em the AI tools and then instead of taking six months to deploy that feature, we'll deploy it in six weeks.

[00:24:50] Barath: Absolutely.

[00:24:51] Josh: And he said, we chose that path. And he said the reason was because they had the conversation that said,

[00:25:01] Josh: of all of our competitors, somebody's gonna choose that path

[00:25:05] Barath: True,

[00:25:06] Josh: and then they will be the fastest to market. And if they are consistently the fastest to market, we will lose the ability to win deals and we will become irrelevant. And he said, so we may be super efficient, but we're irrelevant. We're outta business.

[00:25:24] Barath: I

[00:25:24] Josh: And he said, so we chose that path because somebody's going to. And I thought that was a really interesting take that quite honestly, I hadn't heard anybody talk about that perspective before.

[00:25:34] Barath: Actually, Josh, that is the to be open. Most of the institutions are about reducing cost. They want to become more efficient. For sure. Let's take the same analogy, what you mentioned. There's a 10 member team delivering products over six months. They want to make it efficient of 10 member team coming down to two or three, but reinvest the seven to do more.

[00:26:04] Barath: They're not reducing and letting go of the seven member team. They're reinvesting the seven member team to actually deliver more products, more work and supporting the business initiatives and coming out with lot more. improved customer experience or improved revenue, or improved efficiency. now, depending on the customer's maturity of their existing legacy estate versus a FinTech or a product firm, the fintechs and product firms are able to move much faster, right?

[00:26:40] Barath: And their ability to adopt some of the tools, things that are happening. Like a client of mine in the insurance sector two weeks he and his entire team spent almost a week in San Francisco and he comes back so excited about what he saw and he tells me, ETH, we are not doing anything that what I saw in San Francisco, the potential possibilities are humongous in terms of what we can adopt and the throughput of productivity can be Forex.

[00:27:12] Barath: So his point was not to reduce, his point was to increase the throughput productivity velocity with better quality and do more for his business. That is consistent, Josh, across the board where everyone wants to

[00:27:30] Barath: do more with the kind of effort you have.

[00:27:35] Josh: Are you seeing any differences across.you mentioned like the maturity of how an organization is leveraging ai. Are you seeing a difference, you said across the different, countries, like the percentage of utilization. are you seeing that maturity and that percentage of utilization, is that also having an impact on, how they're thinking about strategically using these tools?

[00:28:01] Barath: Yes. Again, goes back to the organization's. Ecosystem architecture, techno technology debt. Unfortunately, last four or five years, the organizations have not invested enough in modernizing that technology because of which enough of tech debt started, creating, and now that is becoming an impediment today in terms to move faster in, bringing new products to the market.

[00:28:33] Barath: While you said or the CFO or they are bringing in concept of AI first to everything. But as you percolate to two level three levels below the reality hits on real ability to adopt, that's number one because of the legacy estate that you are living with. The second element, which came up.

[00:28:57] Barath: Even yesterday's dinner meeting is the change management that is required in adoption of ai. First culture is not easy and that is not taking center stage yet. In terms of the new ways of working. Using AI needs you to totally reimagine how you should be operating in new scheme of things. Like our chairman mentioned this to a client sometime in early this year, Feb, March time period of this year.

[00:29:33] Barath: He gave a perspective that an enterprise, a large enterprise, if you want to radically simplify and, improve your cost base to reinvest, the way to look an enterprise level is imagine the enterprise operating with agent, AI architecture. Right either on the process or on the technology side.

[00:29:58] Barath: Apply agents and look at what agents can really perform the process that is being done today manually or by different systems. What will cannot be completed by the agents for various reasons. The residual is what the manual effort should be thought through rather than taking a process and trying to digitize you.

[00:30:23] Barath: Apply agents first, right? And then look at what is the residual real manual effort that you need to, have in place for execution or STP. Now, the same thought I took with another partner in one of the big four firms, and he and I were deliberating this and it actually resonated. He was speaking to one of the global banks in us, right?

[00:30:48] Barath: And they were all. Moving towards use case model. They were identifying 20, 30 use cases and they wanted to apply, AI to those use cases. But this gentleman was that if you have 40,000 people doing operations in various form and shape, it is not about use cases. It is about agents literally coming in and, looking at the entire process architecture and business architecture see what, how does the agents fit into that.

[00:31:24] Barath: But have people adopted to execution to this extent? No, not at all. Far away from implementing Totally. An agent architecture but in a midsize to fintechsare much ahead if you ask me. And we cater to fintechs a lot in my own space. I serve almost 45% of my business to fintechs,and then we see a huge adoption of, AI tools and processes, which are much faster.

[00:31:56] Josh: yeah. you brought

[00:31:57] Josh: up a really important point,at the start of that, which was the tech debt. And that I, man, I think that was a super true statement, right? A lot of times and no fault of their own, right. A lot of times our senior leadership team we're so far removed sometimes from the actual process or what's, driving or doing the work. And we say, okay, we have this new AI tool. Go add it,

[00:32:24] Josh: go make this faster, right? And then when it actually goes to be implemented, they're like, that doesn't really work in our tech stack. Like we're operating on whatever old piece of technology that needs to be updated first before we can actually maximize the benefits of it.

[00:32:41] Josh: And so I think one of the things I've heard a lot of times as I talk to different folks is what they're seeing is. And, just call it what it is. A lot of times in the sales presentation, the sales team's Hey, buy this big super cool AI piece of software. We're gonna come revolutionize your business, we're gonna make whatever it is, right?

[00:33:00] Josh: your loan portfolio, 96% more,operationally efficient and ah, it's amazing. And then I ask him, I'm like, okay, so you implemented it. I'm like, we implemented parts of it. We couldn't implement all of it because of this, that, or the other. I'm like, okay, then what is the efficiency gain that you're seeing?

[00:33:16] Josh: And I'm like, it was like 4%. I'm like, so it wasn't the 96 that you were promised, right? I'm using a grotesque example here, To make a point. But, but I think the point stands is that I think a lot of people are struggling to see the full expected value of it. But I think where this gets really interesting is, I'll credit, I was talking to a customer of ours, Joey Ruel, the CIO of Central Laic Credit Union, who is doing a lot of cool stuff with his credit union and ai, like I would argue he's very tip of the spear doing some really cool stuff. And he told me, he said, one of the, one of the things that he has had to really like reframe with a thought process in his organization is you have to think about traditional software and AI differently.

[00:34:04] Josh: And he said, with ai, it's the first time that you can eat the cake while it's still baking. And he said, just because it's not complete or perfect, doesn't make it useless. And his point was, okay, if the promise was 96%, but you're only getting four, is the four still worth it? And if there's enough value still in the 4%, then it's better than 0%.

[00:34:32] Josh: It's better than doing it completely manually the way you were doing it before. And again, I thought that was a pretty cool perspective to take of, hey, this is a balance. And his other point was he said, it may be 4% today. And he said, this stuff is evolving so stinking fast. He goes, tomorrow it might be six and next week it might be 12, and then six months from now it might actually be the 96%.

[00:34:56] Josh: And he's, and we're also just not used to that in traditional software. Like it just doesn't move that

[00:35:00] Josh: fast.

[00:35:01] Barath: I think that's a beautiful point. What you're bringing in. I was almost going to, the thought, What you're saying, but you completed it. What matters incremental progress we are able to show rather than not doing anything. The good thing that we are observing, Josh, is none of the customers that we are engaging with her in a state of not doing anything.

[00:35:23] Barath: Almost, at least I like, if I take last six months, I've connected with a hundred plus clients, right? That is average that. in my top customers speak or connect, none of them are in a state of nothing or be cautious.Everybody's in the journey. Now, the journey varies in terms of pilot or production or somewhere in the middle.

[00:35:47] Barath: as you called out outright. Most of them have not realized we truly believe, 25 to 30% engineering productivity is possible. Most of them have not realized, but with lot of product that is existing in the market today, they already are realizing anywhere between six to 15%. And for a scale of organizations spending billion dollar it budget, six to 15% is a very good starting point, no

[00:36:16] Josh: That's a big number.

[00:36:18] Barath: And of course they want to do more and there is potential to do more for sure. Butas, as I said, the legacy becomes such an important element. And when you say legacy, more important than legacy is not only about the technology debt, it's the data. Because anything on a See AI from persistent angle has three dimensions.

[00:36:42] Barath: AI for software engineering, that is the first dimension that we call out, which is where your productivity comes into play on your software development. So AI for software engineering then comes a dimension of AI for data.

[00:36:57] Josh: Okay.

[00:36:58] Barath: data itself is a fundamental foundation pillar, and if you have not sorted out your technology, debt or legacy in terms of data, then you cannot apply data, AI for your customer insights, decision making faster, processing, impossible,because AI is already big challenge in terms of hallucination, in terms of.

[00:37:22] Barath: Various other challenges, it, is still being solved for. And if you feed half baked or half baked data, which is siloed in different places, then by default you are actually going to get an outcome, which is not appropriate and you're taking a huge risk. And that is where preparing the data in all form and shape to feed into AI becomes a crucial element.

[00:37:50] Barath: And that is where the legacy modernization on data becomes very important, right? So that's the second dimension of ai. The third dimension of AI is process. How you apply agent AI solutions for digitizing the processes and the data is more of a foundation for both insights and process digitization.

[00:38:14] Barath: That is how we see the three elements of, AI in, in our lifecycle assets, of how we engage with AI for engineering, AI for data AI for process.

[00:38:25] Josh: that makes a lot of sense. I like the simplification of thinking about it.I'm glad you brought up the data element because that is an important element, right?

[00:38:33] Josh: it's only as good as you feed the data into it in, and the more or less complete, the better structured it is, the easier it is to actually comprehend versus make guesses.

[00:38:45] Josh: has a huge impact. and I think this is one of the things too that I, just bringing it back to a very simple state, this is why it's so incredibly important to maintain having an expert in the loop in these types of tools is, man, I tell you what bro. they make you feel, pretty smart. until you realize something's not right and you didn't know

[00:39:05] Barath: True.

[00:39:06] Josh: You know what I mean? And so I think a part of it is, a big part of that is feeding the right data in so that you get well grounded responses against that data versus making stuff up.

[00:39:20] Josh: 'cause that's a

[00:39:20] Barath: yeah. yeah. And that is where the element of what you and I spoke whole governance and guardrail come into play. And especially when you feed half baked into AI tools and then you use the outcome for some of your decision making. And if you don't have the right governance, that could be hugely disrupting for your brand.

[00:39:44] Barath: I was speaking to one of thecustomers, a month ago. it was actually also another podcast where a client about the whole governance and guardrails, and I really liked what he. But it is not about moving too fast in adoption of AI and getting some benefit of efficiency or revenue.

[00:40:06] Barath: But govern it and with the right guardrails, one single challenge or a problem that hits my customer can derail my entire brand. And that is where it's, even for me as a service who's consulting to my customers, actually makes me think to step back on the ideas and thoughts we are sharing with customers on adoption of AI especially that are customer facing.

[00:40:36] Barath: And you have to be extremely thoughtful about where you're adopting

[00:40:41] Josh: yeah.

[00:40:44] Josh: Uh

[00:40:45] Barath: have to keep ourselves, put ourselves in customer shoes. And thinking, is this solution going to put them into trouble, or maybe less efficiency, but gives them a better journey? I like what you said, and that's what I keep telling my client.

[00:41:01] Barath: Incremental progression, right? Four to six, to eight to 12 to 25 to 30, rather than jumping onto 30% on day one, and then some mistake happening in six, nine months, which takes away the entire

[00:41:15] Barath: credibility or the trust in the system.

[00:41:19] Josh: Yeah, no, I mean I, it's probably an

[00:41:21] Josh: overplayedexample.and it's a simple example but, and I hate to do it to 'em 'cause I drive a Ford truck. But,this was probably, what about a year ago or so, Ford launched an AI chatbot on their corporate public facing website. And you could go in and you could query Hey, I'm looking for a truck that can tow 6,000 pounds but gets 23 miles per gallon.

[00:41:47] Josh: which Ford truck should I get? It worked great at that before. Didn't put some guardrails around it. And you could also ask it, what sucks about Ford? Why should I not buy a Ford? What issues has Ford had with their warranties lately? Why should I buy a Chevy instead? And it would answer 'cause it was just connected to an open chat GPT.

[00:42:09] Josh: And the best part was people quickly learned that Ford essentially had either a massive or an unlimited budget for API tokens

[00:42:19] Josh: and people were going in and running massive projects to go and do things like generate code

[00:42:26] Barath: yeah.

[00:42:27] Josh: in Ford's instance instead of their own, so they didn't have to pay for it.

[00:42:31] Josh: And so there's all these screenshots online of people using Ford's chat bot to, to do all this stuff, right?

[00:42:39] Josh: Or to have it even dog on its own owner. So yeah, you think about, okay, yeah, it probably helped a few people pick the right Ford vehicle or find a great local dealer. But man, how much did that cost them in reputational cost? Now I would argue, I'm also a marketing person, so publicity's, publicity and, even bad publicity can be good publicity, so they got a lot of news out of it, and, maybe people looked at Ford that wouldn't have before, but still, you have to argue.

[00:43:09] Josh: there's a brand and a reputational hit to it. and like I said, that's a super simple example, but a great example of just not having the right governance in

[00:43:16] Barath: No. Another example, which I'm assuming also is overused. I don't know if you've come across this, the whole, challenge that happened on Apple credit card. Whereliterally it's an open case that came out husband and wife had applied fora card with a credit limit. And, it is AI driven solution. And, the husband got higher credit limit credit limit. the partner got a lower credit limit and,actually wife had a better credit score and various other parameters that showed she should be having a much higher credit limit than him. And yeah,the AI solution gave a different, approval process or a credit as a guidance.

[00:44:02] Barath: And then they investigation and it couldn't answer why it and the guard rates were not in place. And that is what becomes, even more important and crucial.

[00:44:14] Josh: Interesting.

[00:44:15] Josh: Yeah,

[00:44:15] Josh: trying to think of two there, man. I wish I could think of the example off the top of my head, but somebody big recently, kudos to them. I think for actually admitting what happened. But somebody big recently had a huge, I don't remember if it was an outage or a massive product issue with a piece of software.

[00:44:34] Josh: And it was because they were just trying to run too fast and they were doing AI code generation without an expert in the loop, without somebody validating it. And they pushed dirty code to production and brought a system down or had a major issue. I can't remember who it was, but it was somebody big recently.

[00:44:49] Josh: and to your point, yes, maybe they jumped straight to 30% increase in engineering efficiency, but all it takes is that one massive instance like that, and you could totally erase that 30% that you'd seen last year with just trying to dig outta that hole. So there's, it's a double edged sword.

[00:45:09] Barath: absolutely.

[00:45:11] Josh: what are you seeing in terms of the differences

[00:45:14] Josh: globally in terms of governance and just how you. Companies and even geographies are thinking about governing utilization of these tools differently.

[00:45:23] Barath: no, clearly Europe in terms of data protection and in terms of how the information has to be leveraged by ai, Europe is one of the most protective,in place, right? Where any data going out is not, it's totally prohibitive and.

[00:45:44] Barath: That can be another reason where it may be moving slow. Some of the firms there may be slow. but the data privacy laws and governance laws in Europe, is and that, has a very good protection the people element is what I, that comes to my mind immediate, but even in us, to what form and shape that matters. And similarly, if you go data is, one of the reasons where many customers started liking the with them is because it's an on-prem solution rather than on cloud.

[00:46:17] Barath: And by default, financial services customers prefer solutions that are on-prem because they don't want their data to go out. And, neither they want to. Have the models leveraging that is passing anywhere outside their ecosystem. So that is also bit of a, again, an adoption of full-fledged AI tools.

[00:46:42] Barath: Most of the AI tools today are on cloud, most of them, right? and that is where some of the challenges and adoption of even the software development life cycle, what is required for customers to full fledge leverage. Starting from requirements to deployment, customers are not able to fully evolve because of, the data privacy and, the data challenges they have, right? Andeveryone can build their ownmodels at a, at an org level. They need to depend on external, external platforms who invest. Billions of dollars in developing right?

[00:47:23] Josh: Yeah. Yeah.

[00:47:25] Josh: something that, can we go down this path? I'm curious what your thoughts

[00:47:28] Josh: are,

[00:47:28] Josh: the sheer volume of cost a associated with this ai boom. I don't even think my little brain has been able to wrap it around. we are talking ordin amounts of money just like we've never seen before. And it's everything from just the cost of the physical hardware to the land real estate to put the,facilities in, to energy consumption, to cooling.

[00:48:06] Josh: and you look at, there's a lot of talk about how even just some of the major platforms like, open AI and what their valuation is, how they're getting funding. there was recently some news about,them even looking for some government assistance, which I don't know, I'll put my personal opinion out there.

[00:48:28] Josh: I think it's scary as all get out.

[00:48:31] Barath:

[00:48:31] Josh: sounds like a, what's

[00:48:33] Barath: and the famous circular reference of investments

[00:48:36] Josh: So do you think that there's also a huge risk of a bubble of some of this stuff? because, sorry,I feel like I'm a little all over the place. I'm gonna pull this back together. But, it was because you were referencing like a lot of think about just a community financial institution in the us right?

[00:48:55] Josh: This is no shade or disrespect. You don't have the money to build your

[00:48:57] Josh: own competitor to open ai, right? we know what this thing costs, right? And so you've gotta rely on some of these big, firms to be able to access some of this. And then you run into the scary risk of, yeah, how much of our data can we safely feed into that?

[00:49:14] Josh: But two, you run into the due diligence risk. what happens if all of a sudden open AI one day says. Hey, I know we sunk trillions into this thing and we said it was worth even more trillions. Turns out nobody else thinks that. And we're outta money and we're done, and we're closing up shop, and you've built entire businesses practices around this, and all of a sudden it's gone Poof, out of the thin air.

[00:49:38] Josh: What happens? I don't know. That was a huge loaded statement. I don't know where you're gonna take this, but run with it.

[00:49:45] Barath: it.

[00:49:46] Barath: is clearly. let's segregate again, going back to what, my let's segregate the firms into,the leaders innovators.people who will follow and people who will leveragealways, right? If you segment the institutions in the industry into these three categories, the big players evolve their own models, right?

[00:50:03] Barath: And they have massive budgets available with them, and they will invest and they will evolve their own models for sure. But the mid-size institutions to small institutions, as you called out,impossible to build their own models, right? That is not what it is. And they have to depend on the hyperscaler models or various other things that is existing in the, what most of the institutions are doing, Josh, is nobody's going to put all their money in one single, model at all, right?

[00:50:37] Barath: the customerssmall language models to models, depending on their size and scale. they're avoiding. Tokenization for every request. and like our own solution actually builds on-prem SLM rather than LLM. And the cost of processing is far lower than what you would see in different, areas, right?

[00:50:59] Barath: So there is an orchestration layer that is coming on top of different,foundation models and the orchestration layer becomes so critical to leverage an open AI or different, model solutions. And even if something goes wrong, for some reason, putting all your eggs on one founda, one specific model.

[00:51:18] Barath: And the orchestration layer is able to leverage what is the best solution, best model for that particular problem statement. And that is how most of the institutions are evolving rather than putting all their money on single page. But yes, the kind of processing that is required. You saw Microsoft has just recently.

[00:51:39] Barath: Launched, the whole new data center in Atlanta, like massive effort. Google has something both in India and in us.The plans are to expand, so data centers and the processing units are going to significantly scale to support. But I also believe Josh, at every point the models will become only better.

[00:52:03] Josh: Like the old Intel chips that, we saw that, it kept on doubling in terms of, processingpower when the size kept on reducing by two. Yeah.

[00:52:14] Barath: Very similar thing is going to happen in the processingAnd,the kind of load it is taking is going to and that will evolve in far more adoption That is my, my, thought process at this stage, I would say. But as you said, it's significantlyevolving. it evolvesevery month or every two, three months. have head my team, he comes out with a white paper, monthly basis. and the kind of brings out to theeven from a firm like Persistent, is significant, right?

[00:52:46] Barath: In terms of software, in terms of the new solutions he brings to the table is significantly high. And now imagine the big players were investing a billion dollar or multi-billion dollar in terms of evolving

[00:53:00] Barath: their models. Like we, we keep seeing that only improving.

[00:53:05] Josh: Do you think the long-term ROI is there for

[00:53:07] Josh: them?

[00:53:09] Barath: I, I can't comment on that at this

[00:53:12] Barath: stage.

[00:53:12] Josh: I know I, yeah.

[00:53:14] Josh: I'm not qualified. I'm qualified to ask the question. I ain't qualified to answer it

[00:53:18] Barath: More importantly,

[00:53:20] Barath: I don't know. I am as an individual. There's another very close friend of mine. compliment each other by default. I'm more optimistic and sometimes that fails me to see the challenges or risks in certain areas, but I only see what is the way to solve a problem, and that makes me move towards more optimism than what could go wrong.

[00:53:44] Barath: And that is what this gentleman, whom I used to call as a chief risk officer, always brings out what all could go wrong. So he and I used to deliberate in, in different, situations, and he brings out 10 things that could go wrong, things, how to solve that, and then it results in a best possible outcome.

[00:54:03] Barath: I, I genuinely don't know Josh at this Things. I, hope for challenge that comes in, there will be a solution that evolves.

[00:54:13] Josh: Yeah, because you touched on it earlier, right? they've gotta, just as we are trying to use the tools to become more efficient ourselves, I think at the same time they've gotta find ways to become more efficient themselves.

[00:54:25] Barath: It'll

[00:54:25] Josh: Otherwise, we just, we can't, I, I don't think it's sustainable to keep sinking trillions of dollars into this infrastructure and these companies.

[00:54:35] Josh: And then, at a certain point, if organizations like yours, ours financial institutions are not seeing the 96% that they were promised, and they're not seeing the ROI, maybe they don't completely abandon it, right? But they're gonna start to say, Hey, this maybe isn't worth what I thought it was worth this is I'm not willing to pay what I thought I was willing to pay.

[00:54:58] Josh: And so either the productivity gain has to come up, the cost has to come down, or some variation of the two.

[00:55:05] Barath: the scope is huge for adoption.

[00:55:08] Josh: the one of the recent podcast I heard, somebody was saying industry is what, 5 trillion in terms of services. And the current business case that is being thought through by Silicon Valley firms or firms investing in AI is assume 10% of that market we are able to disrupt.

[00:55:30] Barath: And that is huge business. Even only 10% of the services market able to disrupt. And that is what everybody's investing on, on, on the AI bandwagon today. Now is it overheated? Arguably, yes, for sure. But as there could be a challenge of coming down, but I think people will find solution and it'll only improve and go up

[00:55:58] Josh: Yeah. Yeah,

[00:55:59] Barath: II was, hearing, this gentleman who is the professor in the Polishand specializesAI safety.

[00:56:06] Josh: Okay.

[00:56:07] Barath: was the super podcast that but also made me reallyunderstand how things are evolving. If by 2030 he believes that even plumbers and electricians, you will see AI robo coming and, disrupting those jobs.

[00:56:24] Barath: If that is the level thinking that is happening in the industry with, not just in the services space, also in the physical space,imagine the amount of learning, processing that you need to do for that volume of work. So the scope is huge, and that is where all these investments, I hopewill translate into something tangible. the point is always, things move much faster at the beginning than what you can consume. So it'll slow down, it'll dip to a bit, and then people will course correct and then move up.

[00:57:07] Josh: Yeah, I agree. I, I think that's one of the things I find really interesting too, is you just look atthe different acceptable adoptions of different people based on personalities and things, right? And,everything is great in

[00:57:19] Josh: theory. And so yeah, we can talk about how by 2030 we're gonna replace plumbers with all of this.

[00:57:26] Josh: Will it actually, I don't know. maybe it doesn't. but I think it's funny. what was it? One of the new,oh gosh, again, man, my brain is just failing me today. But, somebody just released the first like h home assistant robot that will do laundry and dishes and all of that, and I think it's what, 30 or $40,000?

[00:57:46] Josh: it's not cheap, but it's not like completely unrealistically out of reach for everyone. It's not it's like $10 million. And I told my wife, I was like, I totally want one. And she was like, no way in hell is

[00:58:01] Barath: too expensive to start with. I,

[00:58:03] Josh: she was like, oh no. It wasn't even about the money for her.

[00:58:05] Josh: She was like, no way in hell is

[00:58:07] Josh: that thing coming into our house. She's not a chance. And I was like, what are you talking about? I was like, how much freaking laundry do you do a week for two toddlers? I was like, it's insane. I was like, imagine if you never had to do that again, and you had a robot that did it for you.

[00:58:20] Josh: And she's no, I will do the laundry. There's no way that thing's coming in our house. So it's like it even exists. It's like I could literally go out and purchase one if I had the money

[00:58:32] Josh: and have it, but the actual adoption of it is different. I use a little

[00:58:38] Barath: to that

[00:58:39] Josh: What's that go? Go

[00:58:40] Barath: for a minute. Stick to that one specific LAR use casefor that robo AI robo to be fully trained and not ending up mistakes at home. Imagine the amount of training and data required and processing required to do

[00:59:00] Barath: that, and now look at the kind of investment that's happening.

[00:59:05] Barath: Maybe the investments is not enough. If you end up in 20 27, 20 30 with a hundred million of these robots that is required globally, then comes the affordability. Affordability will only keep improving in terms of what is costing 40,000

[00:59:25] Barath: today. I'm assuming it'll keep reducing significantly over that period of time.

[00:59:30] Josh: Yeah. super

[00:59:31] Josh: simple example. This was really funny actually. I don't know what made me think, even look at this. we went to Costco the other day just to do the, by week or the twice a week, Costco run or the every two weeks Costco run, right? And we walked in and, Costco always has the flat screen TVs is like the first thing you see when you walk in.

[00:59:50] Josh: And I never really pay attention. We've ha our tv, I don't even know how old it is at this point. But it works just fine. there's nothing wrong with it. I don't need, we don't even watch that much tv, so I never pay attention for whatever reason, far off this time, I paid attention. They had a 95 inch ED

[01:00:12] Josh: TV that's like as thin as my credit card for under a thousand dollars.

[01:00:19] Barath: Oh, a hundred thousand dollars. Wow. Okay.

[01:00:21] Josh: Under a thousand dollars. I was like, this is mind boggling.

[01:00:26] Barath: ah,

[01:00:27] Josh: I

[01:00:27] Josh: was like, could you imagine, like even just going back five

[01:00:32] Josh: years ago and telling five years ago, me,actually, no. I didn't even think about this. So funny story about rtv. So when we moved into this house, actually almost 10 years ago to the day now, that's kinda cool.

[01:00:48] Josh: when we moved into this house. We bought a new TV for this space and we bought it from Costco. And I wanna say, whatever, we got like a 55 inch or something like that. That was like the flat screen. And I wanna say we spent two grand on it. We're like, oh this is really cool. We're like, we're buying a really nice tv.

[01:01:07] Josh: This is our first like big nice TV for this house, right? Which now is like a joke. But what happened was we bought it through Costco, we bought the warranty,

[01:01:17] Barath: oh.

[01:01:18] Josh: we bought a three year warranty on it. Two years into owning it, it went kaput. We went back to Costco and we returned the TV and I don't even know what it is now.

[01:01:35] Josh: So I'll probably get, I'm not gonna use the numbers 'cause I'll get it wrong. But let's say we had a 55, now we have a 75. It's a far better tv. But the craziest part about this whole thing, Barra. Was now we have an exponentially nicer TV in that two years. went from looking nice in the space to completely filling the space and, but the craziest part was, I wanna say we got 500 bucks back to use it Costco on groceries. So we got a way nicer TV and a refund

[01:02:09] Barath: See, that's

[01:02:10] Barath: the whole point.

[01:02:11] Barath: Yeah. That's the whole point I'm referring to.

[01:02:13] Josh: Yep. And that's exactly what you're

[01:02:14] Josh: talking about

[01:02:15] Barath: I am an electronic gadget, addict, right? So for me,I change my televisions every or two years. I keep changing. So if a 95 inch televisionfor less than thousand bucks, when I land in New York, I'm

[01:02:27] Barath: going to

[01:02:27] Josh: Yeah. You're gonna Costco?

[01:02:30] Barath: absolutely. I'm a, that's one addiction that sometimes a bit

[01:02:34] Josh: Really? Oh, that's funny.

[01:02:36] Barath: oh, yeah. I'm literally a huge Apple fan. My, all my products around me are Appleand I keep changing it. Every, I

[01:02:44] Barath: don't

[01:02:44] Josh: Okay.I have to ask then, 'cause I've noticed as

[01:02:47] Josh: your, do you really have two Apple watches on?

[01:02:50] Barath: no. One Apple Watch. Whoop.

[01:02:51] Josh: Oh, okay. Yeah. I was like, that's hilarious. That I was like, that might have been a first. Maybe you have the work phone and the personal phone with Apple.

[01:02:58] Barath: no, I Phone. I have multiple phones, but no onefor sure. I'm using Yeah. Yeah. So I'm

[01:03:05] Barath: a

[01:03:05] Barath: big electronic gadget stuff. Yeah. Which is using AInow.

[01:03:08] Josh: yeah.

[01:03:09] Josh: everything, right? And I think that's to your point too, like the scale is going up pretty significantly. But I do,I, I genuinely have a personal, concern for is. Is the utilization ramp the same as the cost and the governance that needs to go into

[01:03:28] Josh: it? let's

[01:03:29] Barath: the, yeah,yeah,

[01:03:31] Josh: open AI has clearly

[01:03:33] Josh: proven that they will not treat your data with respect.

[01:03:37] Barath: yeah.

[01:03:38] Barath: We already know this. Like this is not even being hidden,See that still keep bringing back reference of my chairman,in pockets because he's a technology visionary. What I've seen,in an individual. And he made the statement, this is 2025. he is addressing persistence 450 memberin the forum.

[01:04:02] Barath: And our CEO requested him to give a perspective of the technologyevolution. And while he gave lot of futuristic statements on what he sees a disruption in engineering agent pki, he spoke about it in Jan 2025 adoption. But one thing he called out was the governance and risk the entire,

[01:04:26] Barath: and he even said, which I still believe today, after 11 months,it is not fully evolved in terms of what is the right way to govern.

[01:04:35] Barath: And what is the right way to manage the risk in terms of what is getting produced or the models coming out

[01:04:44] Josh: Yeah.

[01:04:44] Barath: right now? if we take our own platform on the AI for engineering, it has a combinationof generative AI for developing the code, but it is also complimented by what we call deterministic AI for validating what got generated with certain guardrails, certain benchmarks, certain kind of quality, parameters to see what got created is the right, code base or data points that need to so there is a maker and checker

[01:05:15] Josh: Yeah,

[01:05:16] Barath: in the solution that needs to be thought through but is it entirely foolproof?

[01:05:21] Barath: Not it. And that is whereAnd the guardrail and the whole element of I was speaking to another, author a week ago on the whole Reliable AIand both of them have to go hand in hand.

[01:05:34] Josh: yeah.I,

[01:05:36] Josh: you brought up a really important point,a second ago too, of, what's the old saying? Fortune favors the bold, right? At a certain point, yes, there's risk, there's governance that needs to be taken, but like we were talking about earlier, like somebody is going to do it.

[01:05:53] Barath: Yeah. yeah.

[01:05:55] Josh: So some people will fail. Some people will learn some hard lessons and some people are gonna do some really cool stuff. And

[01:06:03] Josh: the people that do some really cool stuff are the ones that are gonna write the history books right now. Are all of us gonna be the ones that do the super cool stuff that writes the history books?

[01:06:14] Josh: No, absolutely not. But like you were talking about too,it's, especially thinking about it at, say, your institution, how fast of a follower are we comfortable with being there? And what is the risk if we don't follow fast enough?

[01:06:28] Barath: Oh yeah, that's absolutely See beyond the point, your questioned. And we have seen enough examples in the last 50 years where if yourself, if you don't reimagine yourself, Yeah, The existence is absolutely question. And that is where be. It's a fine balance. I agree. if you don't reimagine yourself, yeah. Come another five, 10 years, and it may not be even five, 10 years, it may be three to five I like what you said, earlier. A 10 year plan. Oh,

[01:07:01] Josh: Oh, yeah,

[01:07:02] Barath: spoke about this before thepodcast, right? I don't know who asks a five, 10 year

[01:07:07] Barath: plan.

[01:07:07] Barath: yeah. No, the joke he's referring to is that,I won't name any names, but we recently responded to an RFP that asked for our 10 year roadmap, and I went over half jokingly to our product management team. I was like, how do you guys want us to answer this? And they literally threw their arms up and they start laughing.

[01:07:27] Josh: They're like, I don't know, if we know what we're gonna do in 10 years from now, they're like, we should be fired. They're like, 'cause all that means is it'll take us 10 years to do anything. They were like, ask me what my one year roadmap is. That's about it. They're like, technology

[01:07:41] Barath: it's one to three.

[01:07:42] Barath: Yeah.

[01:07:42] Josh: fast.

[01:07:43] Josh: we need strategic direction that goes

[01:07:45] Josh: further than one year.But in terms of committed roadmap things, unless there's certain projects where,we used, we use the example all the time of Fed Now, right? And Fed now took longer. But why? because there was dependencies.

[01:08:02] Josh: Like we had to have the Federal Reserve actually release Fed now, so that's why it took a while. So there's things like that, but man, if it's solely in our control. And we're saying, Hey, this is gonna have to take us 10 years to build, man, it better be something pretty big. 'cause that is just not the world we live in anymore.

[01:08:21] Josh: and that is, I think, a big change. I think especially for,very legacy industries like financial services. For all intents and purposes, a lot of elements of financial services have changed very slow over the years. and now we're seeing a rapid pace of evolution, unlike we've ever really seen

[01:08:39] Barath: cool. I was listening to Satya Nadella recently in a podcast, and he made a statement that actually he was actually stating that you have to be looking at 10 to 20 years future in terms of how would It was not roadmap for sure. It was more around, as a firm, what do you want to evolve yourself as a vision, as a entire journey, which is absolutely needed.

[01:09:07] Barath: And I'm glad the way he's thinking, in terms of, how they're actuallyleading the journey terms of, adoption of AI through various differentso important to have a vision for yeah, from a roadmap and plan,

[01:09:22] Barath: yeah,you need to revisit that every two, three years.

[01:09:26] Josh: Yeah, your plan has to be ready to make a new plan.

[01:09:30] Barath: That's, That's, my oversimplification, but Yeah. No,

[01:09:34] Josh: it's interesting broth because I would say that's a very common trend. actually, fun fact, you'll appreciate this. going back to how fast some of these, AI systems have evolved. if I had tried this even about six months ago or so, I don't think it would've worked very well. but about two weeks ago, given some of the latest updates that Google AI Studio has brought about, in Notebook, lm, I'm a huge user of Notebook, lm, big

[01:10:03] Josh: fan, and, notebook. LM now supports I think a little over 300 large complex documents that can be uploaded into a single prompt and analyzed.

[01:10:16] Barath: Yes. Yeah, I do.

[01:10:17] Josh: you know what I did? I loaded

[01:10:21] Josh: the transcript of every single episode we've ever had on the podcast.

[01:10:27] Barath: Okay.

[01:10:27] Josh: If you listen from episode one to the end of your episode. I think today it takes a little over nine, full days straight without stopping, right? So it's that much content uploaded into Notebook, lm, and I've started analyzing the podcast through Notebook lm, and it has been fascinating

[01:10:52] Barath: it's crazy how accurate it is.

[01:10:55] Josh: Where I'm going with that is, one of the things that it picked up on as a common trend was maintaining relevancy

[01:11:04] Josh: in the face of disruption.

[01:11:06] Barath: Correct.

[01:11:06] Josh: And to your point, we have so many examples from so many industries that, grotesquely overplayed. One is,blockbuster and Netflix, right?

[01:11:16] Josh: But someone is going to disrupt your business. Period. End of story. If your business makes money, someone wants to disrupt it and make that money in front of you. And there is a lot of money to be made in financial services. Let's be honest. So you are crazy to think that somebody is not looking at that and going, I could disrupt that.

[01:11:35] Josh: How could I disrupt that? And to your point, like our strategic vision looking 5, 10, 20 years out now needs to be, if I was not here and I was somebody who's trying to disrupt what I'm doing, what would I do and how do I need to set us up to head in that direction so that we are in disruptable because we disrupted ourselves.

[01:12:01] Josh: And it was just fascinating to pick that up as a very central common theme throughout 140 episodes of the podcast. It's crazy.

[01:12:10] Barath: it's funny you said that, I think 1 26 episodes are panel and 40 episodes,so I saw

[01:12:14] Josh: I think this one's gonna be close to

[01:12:16] Josh: 140.

[01:12:16] Barath: Okay. Okay. I saw 1 26 when I was looking at it, but not for this podcast. I was preparing for a different podcast two weeks ago, and I wanted to actually understand I didn't get time to go over the previous podcast this gentleman has done in the past.

[01:12:34] Barath: All I did was asking Gemini, okay, help me understand all the podcasts this person has done in the last 12 months, and gimme a summary view one, I could have gone and listened to six or seven different podcasts, but the way it summarized me, all the eight podcasts the person has done in the last 12 months and help me to prepare for my next podcast was like, fabulous.

[01:13:01] Barath: Imagine what could have taken easily five, six hours for me was a 30 minutes engagement with the LLM, the chat, and then I could easily prepare myself for productivitythat it gave me was fascinating. There's so

[01:13:16] Barath: superb examples.

[01:13:18] Josh: Yeah. and there's so many low hanging fruit

[01:13:20] Josh: runs, And I think that's the, I wanna come back to that, but I want to use an example of this exact episode, right? literally Barath and I were joking about, one of the things that I started doing quite a while ago now is, anytime I have, a guest on the podcast, we like to try and do, 15, 20 minute discovery call before the actual recording.

[01:13:41] Josh: And sometimes this takes place. when did you and I do the official discovery call? It was weeks ago,

[01:13:47] Josh: I was in an airport, if I remember right. I was in an airport, like standing in front of my

[01:13:53] Josh: gate.

[01:13:54] Josh: now I have an AI note taker that runs on all of my calls. It took the notes of the calls as soon as we ended the call. I just went to my note taker. I said, summarize our call. Tell me about the guest. Tell me about the topics that we're gonna discuss, highlight, bold, and put them in a bullet point format for me. And then I just copy and paste that. I put it into the calendar invite for the actual recording today. So when Barath and I went to join to actually record the episode today, we could go back and just literally as we pull up the calendar invite, look and see, oh, okay, these were the six topics that at a high level we said, we'll try and touch on, or we'll think about to help guide the conversation.

[01:14:40] Josh: And especially for something like this podcast where it's totally unscripted, right? That just gives us a good guiding light. And that took me nothing, no time. I did that in an airport. Literally, I got off the phone with broth and I was like, all right, I like, they're calling my boarding group. And before I was on the plane calendar invite was.

[01:15:00] Josh: Meeting summary done, bullet points. And now, we're able to come in weeks later and have all the context really beautifully summarized for us going into the call, right? So there's just, there's super simple low hanging fruit, and I think that's one of the things that I've found the most interesting to me over the last little bit, is just examples of how we can find simple, super cool little use cases of productivity gain even in our personal lives.

[01:15:26] Josh: that's pretty neat.

[01:15:28] Barath: agree.

[01:15:29] Josh: Yeah. broth, I,I want to respect your time and your world traveling and, give you an opportunity to maybe,

[01:15:37] Barath: soon as I finish this, I'm heading to the airport.

[01:15:41] Josh: There you go. See, man, the world we live in. Seriously, this has been an absolute pleasure chatting with you. I really appreciate the perspective and kind of the, I don't feel like we went crazy deep into one certain area, but we're able to touch on a lot of different facets of how, globally businesses are thinking about kind of the next evolution of technology.

[01:16:02] Josh: So I, I really appreciate you taking time outta your busy schedule to join me. And, before I let you go though, I have two final questions for you. So first, where do you go to get information about what's happening in your space?

[01:16:15] Barath: Ah, okay. I'm a lot more into visual medium now than, bookreading. So I'm logged into podcasts, various podcasts and varied subjectsand so that is one medium that I've, really got a lot more deep into. so less of reading, lot more into. video podcast.

[01:16:37] Barath: Significant fact, I've started binge watching so many different podcasts on

[01:16:42] Barath: AI and technology and various areas.

[01:16:45] Barath: So that is something

[01:16:46] Josh: any that stand out to shout out that people should go listen to?

[01:16:49] Barath: Ah, okay. That is this gentleman called Hil Kamath, who is the CEO of Zero, the as a brokerage trading firm in India. And he has been interviewing, he interviewed Sam Altman, he interviewed CEO of Uber. prime Minister of New Zealand, And he's targeting podcast for, people group of 20 to 25. And for, me,listening to this podcast is giving me a very diversified thought process

[01:17:20] Barath: of different people, different ideas, different engagements. So that's been something. And few more. I

[01:17:26] Josh:

[01:17:26] Josh: and you're like

[01:17:26] Barath: 24, right? no, there's the reason I'm watching this, there are a couple of them that I don't remember the people, but like the Polish, professor that there was one with Jeffrey Hilton, that, had done. I don't remember the podcast, name I forgot. butyes, there's,Jeffrey Hilton's almost90 minutes podcast. I, enjoyeda Polish professor, Roman Polski, if I'm not wrong, Dr. Roman. That was really good.recently, Satya Na And, Sam Atman did one podcast that heard.There's so many of them, so many of them that I follow.

[01:18:03] Josh: Yeah. No, this has been fascinating. safe travels, back home and, happy Thanksgiving and, congratulations on the two year anniversary of, a conversation that shaped, a future. hey, thank you Barath for coming and being a guest on the Digital Banking podcast.

[01:18:18] Thank you for listening to the Digital Banking Podcast, powered by Tyfone. Find more episodes on digital banking podcast.com or subscribe on Apple Podcasts or wherever you get your favorite podcasts.