Why AI Starts With Better Data, with Parijat Banerjee.

In the latest episode of Digital Banking Podcast, host Josh DeTar of Tyfone welcomed Parijat Banerjee, Financial Services Global Business Head at LatentView Analytics. The episode centered around how AI depended on strong data, clear process design, and a human-first purpose.

Josh and Parijat started with a simple idea: good technology should help people pay better attention. Parijat argued that listening remained the most useful human skill, and he framed AI as a tool that could remove busy work and make space for real focus. From there, he traced the rise of AI back to falling storage costs, wider access to data, and the shift from rules-based systems to models that learned patterns at scale.

The conversation then moved to what financial institutions had to get right. Parijat explained that poor data still led to poor outcomes, while unified data created a single source of truth and faster decisions. He also noted that banks and credit unions faced tighter limits because they needed accuracy, lineage, and explainable AI. Josh and Parijat closed on a practical point: community institutions could use AI well if they built on trust, local relevance, and clear customer needs.

Transcript

[00:00:00] Parijat Banerjee: What influenced me most was the realization that data tells the truth even when opinions don't. And now we have finally the tools to listen.

And that's what AI is providing to us. I didn't get into data because it was fashionable. I got into data because it was fundamental.

[00:00:22] Intro: The Digital Banking Podcast is powered by Pho. 

[00:00:28] Josh DeTar: Thank you so much for listening to and watching the Digital Banking Podcast. If you enjoy our episodes, please follow us on your favorite podcast platforms, and if you're willing, we would love if you'd take a minute to write us a review. Thank you so much for being a part of the journey that is the Digital Banking Podcast.

One thing you'll never hear on this podcast, sales Pitches except this one, Java for Kids [00:01:00] Coffee. Every bag purchased sends $5 to the Children's Miracle Network Hospitals, so now your morning or afternoon cup can make a real difference. Check out Java for kids.org for more.

 Welcome to another episode of the Digital Banking Podcast. My guest today is Parijat Banerjee, the financial services global business head for LatentView Analytics. 

There seems to be a common trend in guests I have on the podcast who work around big, fast moving tech, that while they work in and love tech, they crave human connection more than anything. And when I asked Parijat what was important to him, he immediately responded with giving other humans the gift of his full and [00:02:00] undivided attention.

He said, we live in a world that's so hyper-focused on multitasking that oftentimes we lose sight of the listening part of connecting with people. But seriously, how many times do we do that too often, whether it's in work or in our personal lives, and many times in our professional lives that leads to trying to offer a solution before we've really heard the problem explained fully.

Now listening leads to solutioning, and that's the special sauce. As a diehard soccer fan and player. Parijat said, in soccer, you're taught to be highly in tune with where the ball is and attentive when it comes to you being hyper present and focused and that the team must play together as a cohesive unit, as a society.

We've also moved away from doing things together. [00:03:00] We may be together, but we aren't together. You know what I mean? So how does this translate to the topics at hand today? Well, spoiler alert, this is yet another episode on AI and other things. But from the perspective of how can AI and modern tech actually help us be more present, more in tune, and more together.

So with that Parijat, welcome to the show, sir.

[00:03:30] Parijat Banerjee: Thank you, Josh. Really loving to be here with you today.

[00:03:35] Josh: You know, it's funny, I'm just gonna, call it out super bluntly. We've been doing this podcast long enough now. It's pretty cool. It's a weird humbling honor, but now I get an influx of people and PR firms and marketing teams that reach out and want to pitch guests to us.

I don't know what the percentage of guests I actually take from people reaching out cold is. But I remember when your [00:04:00] PR firm reached out to me and sent it over and I was like, I don't know, another guest on AI and not necessarily working directly in a credit union or a community bank, like, is this really here?

I don't really remember what it was but something in the email stuck out to me and I was like, ah, it might be worth a 15 minute chat with Parijat and we'll see where this goes. And I'll probably just be really nice and be like, Hey, I don't think this is a great fit.

And about five minutes in I was like, oh my gosh, this is awesome. I can't wait to talk to you. I would love to have you as a guest. And I think it was just because we've had so many conversations about AI lately, but I thought you had a really interesting perspective of what got us to this point.

And you just said a couple of things that I think are gonna be really, really valuable insights for our listeners. so I'm really, really, thankful for this kind of happenstance, meeting and the opportunity to have you on as a guest. You know, I wanna start with something that you said on the discovery call that really stuck [00:05:00] with me, and I want to use this as almost a guiding theme for a little bit of our conversation today.

I kind of alluded to this in your introduction. it's something that, people who've heard me talk about AI have hopefully heard from me too is that I'm a big fan of it. I think there's a lot of things that we can do with it. there's a lot of operational efficiencies that can come out of it.

But more than anything, if it can help us to be more present. Because we're better informed, we're more efficient with mundane tasks and things like that. Then it frees us up for me to actually look you in the eyes and have a conversation with you and just be paying attention. And it's simple, silly little things like just having my AI note taker run in the background.

That happened when we did our discovery call. You and I just talked and I didn't take a single note. My AI note taker took it all, and then I went back and I edit it and was like, Hey, these are things that I remember that I wanted to call out. It [00:06:00] just lets you be more present.

I'm curious, when I was asking you about yourself, you kind of brought that up too, and you were saying, Hey, I feel like we're in a world today where everybody's multitasking, everybody's trying to jump to a solution without actually listening and you were like, no, we're missing something. Like the special trait. The special secret sauce is actually just listening, and that's a super human trait, right?

[00:06:27] Parijat: I think you're spot on, Josh, and again, you framed it very well. I personally think attention is the greatest gift. That any human can give to the other and vice versa. But, before I get into ai, I think we should talk a bit about the base, which is data. Now I started working in data back when it wasn't sexy, right?

When AI lived mostly in research labs and old list programs, and telling someone you worked in analytics earned you puzzled [00:07:00] looks, right? It didn't give you LinkedIn followers. Now my motivation came from sheer curiosity, right? And the belief that data could answer questions we hadn't even learned to ask yet.

Now, in the seventies and eighties. AI meant symbolic reasoning. You had expert systems, you had list code running. And by the way, AI is as back as the seventies and even before, right? It's just in four today because of chat, GPT. But by the nineties we saw SQL databases, they were oap cubes, SaaS, SPS, a bunch of different things that were playing in.

But we were still fighting for storage space. Misers, we were holding floppy desks. I don't know if you remember or even seen one, but then in the 2000 something happened, right? Storage got dirt cheap. The internet made this data floodgates burst open. And you had tools like Python and Hadoop and our, and others that came into play.

And fast [00:08:00] forward today. You have these multiple cloud platforms and you've heard of GPUs from Nvidia and machine learning frameworks like TensorFlow and others that have turned AI into a real time business engine.

[00:08:14] Josh: Now, I've stayed in this field because I've seen it evolve from painstakingly coding rules by hand to systems that started learning patterns on their own.

[00:08:26] Parijat: The right example of you being when we were chatting and your note taker was taking notes, right? It's like watching this shy student in a back row grow up to become a TED Talk speaker, right? And 

[00:08:40] What influenced me most was the realization that data tells the truth even when opinions don't. And now we have finally the tools to listen.

And that's what AI is providing to us. So if I'm being honest, I didn't get into data because it was fashionable. [00:09:00] I got into data because it was fundamental 

[00:09:04] Parijat: and I'm very happy it played out that way. And now that AI and data science, and these are all the front page news, I feel lucky to have had the front row seat to the show all the way from lisp to large language models that we have today.

[00:09:22] Josh: I want to come back to something that you said here in a second, but I want to pick up on, you mentioned AI's been around for a long time. It didn't really reach the ubiquitous of forefront of people's minds until Chachi pt, and it's so fascinating to me how one company with the right idea, with the right execution at the right time can just radically change the course of how things are done.

And you think about it becomes, it literally becomes how people even communicate about it, right? I mean, I made a joke recently in a meeting. Where I asked [00:10:00] everybody, I was like, if I were to just tell you that I was gonna have somebody drive me somewhere and not use my own car, what would I be doing?

And everybody at the table immediate was like, oh, you'd be Ubering. It was like Uber completely changed how we even communicate about that. And same thing, like, I mean, it's just so funny. While I'll talk to people and they don't even say ai, right? If I'm just talking to a neighbor or a friend that maybe doesn't work in tech, they're like, oh yeah, chat GPT, right?

That, because that's AI to them. And so it's interesting how just one big forward facing movement can change everything, but you were talking about, and what I wanted to come back to was the concept of. What really accelerated this was the cost of data storage.

Because for AI to really offer a lot of the power that it brings, you've gotta have the underlying data. And it's funny that you bring that up. You know, I'm not a [00:11:00] technologist by trade. I'm not a computer science graduate like yourself. I'm usually the stupidest guy in the room when it comes to that kind of stuff.

And it was funny, I have this brilliant guy on my team who's doing some really cool work for us. And one of the things that he was doing is building out a whole tool for our company to use and all of this. And he came to me yesterday and this is just so funny how relevant it is. He came to me yesterday. And he wants to do this project, and he needed me to approve some budget for him for it. And I almost fell down laughing because he came and it was just funny how he prefaced it. He put it all together, like the business case, in a paragraph for me. And then we were on the phone talking about it and he was like, yeah, yeah, I need this huge database to store all of this data and do all of this.

And he's kind of setting it up, like I'm expecting him to come to me and say he needs a million dollar budget, or he's trying to give me all this good logic and reasoning why he needs this. And I was like, okay, how much is this thing gonna cost? And he's like, ah, I think about like 25 bucks a year.

[00:12:00] And I just start laughing and I was like, dude, are you really coming to me for approval for $25 a year? Like, this is hilarious. That was literally the amount of storage that he was need. And he was talking about massive amounts of data, right? 

I mean, the storage is so crazy cheap. Like it's not really about the storage. Where the cost is gonna come is in all the API calls for all of the different chat sessions that are gonna happen. Like that's where our cost is going to gonna shift to. But you think about it, it wasn't that long ago where if he had come to me and asked for the storage that he needed, it would've been a very different story. 

[00:12:35] Parijat: I think you're spot on there. I always say that, it's only because data got cheaper. Actually storage got cheaper and you could store more and more data that automatically the entire influx of everything that comes in. Today, I don't think anybody talks about data other than you might hear about data centers. 

These huge mammoth buildings that are being put together with racks and racks of [00:13:00] servers that you would see. But it's been this timeline, post internet where we found this, it's almost like a hockey stick, graph, right? Which has grown over time. The pre 2000, early 2000, you know, 2000.

10 to 15 and then 15 to 20, you would see these are significant eras in some way from a data standpoint and how that grows. An interesting segment that comes to mind is when I was attending the GTC for Nvidia. Jensen made a wonderful comment. He talked about the pyramid of how AI looks into, and interestingly, the base of the pyramid today is not data, it's energy, right?

So you have energy, you have infrastructure, and then it keeps going on all the way up. On the top where you're leveraging data to work on these multiple APIs and applications and structures that is across industries that [00:14:00] run. The understanding is data is ubiquitous. It is there, right? I mean, economic times long back somewhere said, data is the new oil.

I think it's gone way past that. To some extent the leverage of that has been because storage has gone cheaper over the year and it's gonna get more over time. I think eventually we are looking at an AI-based system where we are gonna look at the intelligence that you get. So in it's data to insights and insights to action.

And that's where you drive the intelligence and it's gonna be the intelligence that you get per wattage of energy that we leverage. That's the new framework we are moving towards.

[00:14:41] Josh: I have to apologize in advance. I really wish I could remember who posted this, so I could give you credit. If you're listening and this was your post, please give me a hard time and I'll make sure to credit you for this. 

I wanna say it was literally yesterday, or super recently, I think somewhere in the Midwest. There were gunshots fired at [00:15:00] like a governor's house. Not directed at anyone, but just wanted to make the statement. Thankfully, nobody was hurt. But all that was left was a sign. That said, no data centers.

And the post was really talking about one of the next kind of challenges in this whole world of AI is not even just the compute power or the technology, it's literally the land and the fact that everybody wants more intelligence. They want better personalization in tech. They want all these things, but they don't want the data center in their backyard.

And so politics is literally gonna come into play of where are we gonna place these things? How fast are we gonna allow them to grow and take over? And yeah it'll be less about even just finding the energy or whatever it may be. It's literally [00:16:00] just finding the land to put them, that people are okay with.

Giving up the land to put a big giant data center on. I thought that was interesting.

[00:16:10] Parijat: And Josh, if I may respond to that, see the AI race is on, right? It's like the genie out of the bottle. You cannot put the genie back in. And so technology is almost like that, right? It's out of the bottle. There's no way for you to walk away from it.

The only way to do this is to make sure you have the right guard rate. So that's where politics, that's where policies, that's where countries come into being to do that. But in true sense, the race is on, right? Whether it's us, it's China, it's Europe, it's Asia, India, the race is on and every country in its right state would need to get on the bandwagon.

They're already on that bandwagon, right? And you're trying to find out, and you would see many discussions across the board where there [00:17:00] are billions and trillions being invested, but that's not the money that you see in right now. A lot of it is frontend investing for that matter. But you're right, it comes down to building these data centers, feeding them, the place that you put it.

Because end of the day, AI is only as smart as the data you feed it.

[00:17:20] Josh: Yeah.

[00:17:21] Parijat: And as of today, right? You hear POCs everywhere. I think too many companies are serving it. Almost vending machine snacks when it really needs fine dining, right? And that is gonna take its time to build it and give it so that it can start producing what it can produce.

[00:17:42] Josh: I am glad you came back to that 'cause I wanted to talk about that, but I also want to just kind of put this out there that I wanna come back to too. One of the things that I appreciated about talking to you earlier was, you know, you also have kind of a global perspective of how this is being looked at.

What were you telling me? You gotta fly to Mexico City [00:18:00] and then to Amsterdam, and then to New York. And you made the comment that like across the world, everybody's having similar conversations and similar challenges, maybe with some different nuance, with things like geography or with social norms of that country.

But we're facing a lot of the same challenges. We're having a lot of the same conversations, right? So I wanna kind of come back to what you're seeing from a global perspective. Just going back to, yeah, this was the big thing that I really wanted to talk to you about. 

It starts with the data, but it also starts with kind of the process intelligence. And it really is, it's as simple as garbage in, garbage out. And you know, we see that even in some of my own ways of utilizing different AI tools, whether it's personal or professional. If the data set that I've got available to it is really weak, then what I'm gonna get is really weak responses.

I'll give you a great personal example. I started kind of having a conversation. It started in [00:19:00] chat, GPT, and I recently migrated it over to Claude and found Claude was a lot better at it. Just about my own personal health and fitness. I gave it really weak data and we were just having some conversations about it and ultimately the recommendations, the thoughts, it was pretty surface level, right?

And then I moved it into a protected environment and uploaded a lot of data. Everything from blood work reports to connected it to my smart scale at home, to all sorts of different things. And it was fascinating how good it got. Absolutely incredible and some of the different things that we started to talk about.

And it was just, it was such a black and white example of I was expecting it to give me this revolutionary advice without a lot of data to go off of. And then when I fed at great data, it actually got pretty impressive, right? And you see the same thing in our professional [00:20:00] world too. You think about even just something as simple as like a credit union, rolling out a knowledge base bot and having their frontline teller staff have an AI chat bot that's built for them where they can go in and query, Hey, I've got a member who's asking about x.

I need to know this about the product that we offer. If your dataset isn't that great, then the response isn't gonna be that great. And I think what's challenging is depending on where you fall in the spectrum of the organization, your own interactions with different AI tools, you may think that's gospel.

The answer that you get out is great and you're like this is awesome. And it's totally off base. 'cause it just doesn't have the data set underneath to actually give you the appropriate response that say if you had gone and talked to somebody who'd been at the credit union for 25 years and asked them the same question, they would've give you a far better response.

So it does, it kind of all boils back to the data. And that's why you were saying that was [00:21:00] where kind of the catalyst of this general utilization of ai, especially in business for specific use cases comes down to, well, we got to the point where storage was so cheap, we could dump all the data in that what was coming out was actually meaningful.

[00:21:14] Parijat: So Josh, I'm so happy you ticked and tied this back to where we kind of started, right? Jim Berkson had said, data will talk if you're willing to listen, and I'll take the health piece that you put in flawed and you got the output. I'm gonna take it as a broad maybe an anecdote or a use case that I've seen in real life play out. 

In the healthcare ecosystem, typically, right? You have the payer provider patient, right? I say it's a virtuous and a vicious cycle and it often feels in that ecosystem. These are three separate corners, right? Each holding data hostage in its own silo. So you have claims live with insurers.

You have the clinical insights that's provided by the [00:22:00] EHRs. You have the patient and the health records scattered across places. So we were working with one of the largest non-profit insurers. And to give some context this payer had around two, two and a half million member households.

Their entire ecosystem starting from the CEO all the way down for them to take, action, they had data in silos in different places. And think of it, the entire CXO suite had to take action, looking at hundreds and hundreds of separate dashboards. So if you have to look at 50 dashboards and to come to a decision, that's not easy.

That's very difficult. And you're looking at different metrics. You're looking at finance claims management, right? Just to give you some details. And each one of them was contradicting. So you need to go back and find the source of the data again. It was completely, almost tribal knowledge.

So when we started the [00:23:00] discussion, I realized that the entire CX O suite, what they lacked, what they were having was death by dashboards. What they were lacking was one

[00:23:08] Josh: sorry, but that needs to be a t-shirt, death by dashboards. 

[00:23:12] Parijat: They were lacking one source of truth, right? And we said, look, without no much details, you need to make sure that your me metrics are consistent.

And most importantly, the people finally trust what they are seeing. So we ran a project for them called Cortex. Eventually they got the mission and vision a award for it. But ideally what we wanted to do was the organization should have a single structure and we took out all these hundreds and hundreds of dashboard they had and came down to a few that is enough for the CXO to take 80% of their decisions.

And that was a huge win for them. And that's when they started realizing how important data is, right? And now that they started realizing the importance [00:24:00] of data and they found that there was a base that they could work with. They said, you know, I want to leverage AI full throttle on top of this.

And we said, okay, fair enough. If you want to leverage ai, let's understand where the core structure is. And the core structure, if you think of any payer, is. Say you or me, right? Josh is a member. I wanna make sure I have an entire DNA for Josh and not the biological DNAA data, DNA from Josh. How can I build that?

They started thinking through that. It started with around 600 member attributes. Eventually went to thousands and more. But the whole idea was to give an X-ray vision to make sure you can touch and feel all the data components together for jaws. Just like your story in Claude, when you put in all the data in, it starts really working through, and it did, right?

I'll give you an example where it did. Imagine a marketing or sales leader who wants to run a campaign for a [00:25:00] specific demographic of people. Just say you know, wants to look at who's pre-diabetic or has a history of cardiac issues and you know, he wants to go ahead and run this.

It normally takes around three to four weeks before they can get the data. What we did was we started working to build this member DNA AI platform, and it came up where you can ask direct questions like how many members have prescribed diabetes medication or how many members have total medication expenses above a hundred k.

Again, getting this data, which took around three to four weeks across salespeople and the team and the clinical data and all of that together, this unification got it down to 48 hours. So you can ask, put in a question and get that answer in 48 hours. It's very difficult in a broad ecosystem because you have so many siloed structure.

So again, it speaks to. The [00:26:00] takeaway is when you unify data, when you embed AI deeply, you don't just make analytics better, you make the entire ecosystem better, right? I think, that exactly speaks to, what you saw in your personal space.

[00:26:14] Josh: You know, it really does start with the data and the data structure and what's being fed in, in terms of attributes, because one key metric missing can make all the difference in the value of the analysis that comes as an output, right? So, take my personal example, the health and fitness thing.

Let's say I go in there and I tell it that I am 200 pounds. And it makes an assumption that I am six feet tall, that's probably a reasonable weight for me, right? So maybe it's got some certain recommendations for that. What happens if I'm four foot nine or I'm seven feet tall, right? Completely changes that equation.

And so if that data point is [00:27:00] missing and an assumption is made, we could be going down a totally incorrect path, right? Same thing goes for our business decisions. If I'm missing one critical piece of data then the assumptions may be made on my behalf that are incorrect. And so, there's a part is feeding in the right data, having the right data in the first place.

And having a process to identify is the correct data set being leveraged to make sure that my output is trustable?

[00:27:35] Parijat: That is an amazingly rich observation, Josh. I always feel when answers are cheap and abundant, good questions become the new scarcity. And, those are the right questions when we think across the data structured and platform. If I take this broadly across industries, right, and, and maybe let's go to financial services.

Financial [00:28:00] services has been an industry where technology has always been in the forefront. You know, when we had the ATM machines started where you can dispense cash and, you know, people don't need to go to branches, you had the drive-throughs built where you can put in your checks and the vacuum tubes.

Financial services always has been in the forefront. However the data. In financial services, even in huge bands today still sits in mainframes, right? We are in a cloud era. There are only few bands that have moved completely to the cloud, but that data still sits in mainframe.

So what happens is there is a huge silo in data structures. And that's where you see a huge difference between the traditional banks and the fintechs, right? The fintechs are forward looking, they have quicker formats. Cloud [00:29:00] data leverages faster. Whereas if you look at the traditional banks, it is slow moving only because in the structure in which the data is now that that will change and it is changing.

But I think the important part also to notice from a financial services perspective is. It is the most, I wouldn't say most, but yeah, to some extent probably the most regulated industry, right? Other than healthcare and life sciences. And because of the regulations, the need for data accuracy becomes important.

So the quality of the data, the lineage of the data. I remember when 2008, the market crashed and post that I was with a bank setting up Basel, Zika do Frank, all these various focus on risk programs. one of the one of the analysts from SEC and, and an officer from Fed had asked me we were doing this tier one, tier two capital calculations.

And they said, can you take this [00:30:00] calculation data back to the piece of paper? And that was a light bulb moment for me because we massage the data so much, at least in financial services till you come to the final outcome. It shows the importance of the lineage so that regulations can make sure that the data is correct and your outcome is correct. 

I think AI is probably a bit, or the leverage of AI is gonna be. A bit more cautioned in financial services only because you cannot have a black box where regulations are there. You have to be able to explain every piece and activity that is happening.

So it is moving, it's moving fast in some places and slow in the others. It needs to have a very strong foundation of data that is rightfully leveraged to, get our outcomes in ai.

[00:30:54] Josh: That makes a lot of sense. It makes me think of something else that you were telling me that [00:31:00] may at first kind of seemed totally unrelated, but I think is really related to this topic. You were telling me you'd spent some time with GE and working on airplane jet engines. And you realized in that space that Six Sigma was not good enough. While that is a really high degree of quality control in something like a jet engine, six Sigma means we're losing lives. And that's a pretty big deal. And that we needed to get more to 11 Sigma to save lives.

And while financial services may not be jet engines, and if a part fails, that thing falls outta the sky and we lose lives. And that's a really big deal. It is a highly regulated industry, and so we do have to be very cognizant of the quality control levels that we do. But [00:32:00] at the same time, it's gonna move really fast in some areas and it's gonna move really slow in other areas.

Can I put some words in your mouth, and I'd love to get your take on this. Where it's probably moving a lot slower are the areas where we need 11 Sigma where it can move a lot faster is the areas where Six Sigma will do just fine. Because in ai, if it gets it a little bit wrong in this area, it's not catastrophic.

If it gets a little bit wrong in this area, it's catastrophic. And I really feel for the highly regulated industries, because I think what it does a lot of times is that we apply the, Everything has to be 11 Sigma. We can't get anything to 11 Sigma, so we're gonna do nothing. And that's a bummer because you're missing out on the opportunities where Six Sigma would've been just fine and you can do it to that level today.

And so you're missing out on potentially some operational efficiencies or some data intelligence that you could have gathered [00:33:00] because we could have moved fast in this area. But that kind of overarching fear of regulatory compliance makes us take a pause. What are your thoughts?

[00:33:10] Parijat: And I, I'll answer it through a slightly different lens. Whenever I think of a solution, it always has two doors. It has the process door and it has the data door. Now, the process door always defines what the typical structure process would be, and where you have non-value ads as part of the process and you run with the data door helps you to define the outcomes. 

In financial services, as you rightly alluded, there are some areas, for example, the front end of reaching out to clients, the retail part of it, the marketing, the all of that has much lesser segments for us to be able to leverage AI and work with it, right? But the moment we are getting into [00:34:00] models, you know, we are getting into fraud.

You're looking at, probability of defaults and loss given defaults. It's not that you do not use ai, it needs to be explainable ai. Right? And that is very, very necessary. So, to your point, regulations do slow us. I, they're necessary. Don't hear me wrong. They're necessary, and they're absolutely necessary, at least at an age where the fraud leveraging AI can be really, really high.

People can use that in many, many aspects to go ahead and, and do that at least in anti-money laundering and focus areas of fraud analytics. That's why it's getting bigger and bigger by the day. 

He had observed that computers were adept at tasks that humans considered complex, right? Like chess, math, et cetera. But they struggled to match the perception and motor skills of an infant that was born with [00:35:00] billions of years of human evolution. Now, this is known as the more OFX paradox, and it's extensively used in robotics.

Obviously, more so being leveraged with AI today. But I think it would just as well be applied across industries and organizations, specifically financial services as we drive to gain competitive edge. And thus, to address this scientist in a crib analogy the change in behavior of leveraging the AI.

And leveraging analytics and data and AI as a good to have option towards almost now, using AI as a product and analytics as a regular mission to work with, I think is absolutely critical for future growth of financial services and society as a whole.

[00:35:51] Josh: Why do you think that?

[00:35:54] Parijat: because if you position the amount of uncover that [00:36:00] we can have in areas that we can reach out, so go back to digital transformation. If you look across the world and don't look at first world countries, but look at across the world, you still have more than 30% people who do not have financial access. I think the leverage and use of AI data, AI and technology can help this. And 30% is huge, right? You're looking at nearly two and a half to 3 billion people across the world.

[00:36:31] Josh: Yeah.

[00:36:31] Parijat: I think the leverage of that really helps this scientist in criminology, right? That helps us to provide benefit in the broader ecosystem of society.

And financial services is a key pillar, right? It is a key pillar in society. You did mention about credit unions. I think credit unions play an amazing role in the position of society, if you would, right in some extent. And they have [00:37:00] amazing focus in driving community work across the areas.

It's a different thing that they may not have. The breadth of spend as a large bank does, but there are many positives in which a credit union can start influencing society with the leverage of ai.

[00:37:20] Josh: Yeah. That's a good point. I sit on the board of a small credit union, and actually, literally just last weekend we had our annual strategic planning session with senior leadership. And we all get together in person for a full day session to kind of talk through things.

And Parijat, I just, man, I wish so many people could have just been a fly on the wall to hear the types of conversations. I mean, just hearing the types of things that come out of the mouth of the CEO of our credit union. I'm like, this is why I get excited about this industry. This is why I get passionate about this.

Just to hear them talk [00:38:00] about our business as a business. Look, let's just call it bluntly. The, community financial services, credit unions, and specific our technically non-profit. But if you don't make a profit, you got a business, you can't do the good that you're set out to do, right?

So you have to stay a viable business. It's not about turning a profit, but you still have to pay a profitable business. So to hear them talk about our business as a business, to hear them talk about our customers, the people that we get money from. As members, as people, as humans, not as numbers as people.

To hear our CEO talk about the different segments of people that he wants our credit union to serve and why, and where they need help and support, and they need someone in their corner when it comes to their financial lives. Man, I was fired up. I walked away from being locked inside for 10 hours on probably the first [00:39:00] stunningly beautiful day in the Pacific Northwest, which was a little painful to be inside for.

But I walked away from that going, man, that was a great use of my time. I just, I felt so invigorated from that conversation, and that's a long way of saying I couldn't agree more. I mean, I think the fabric of specifically the US economy is so closely tied to the work in the communities that community financial institutions do.

And you said this earlier, right? AI absolutely is, can be and should be a great equalizer because you're right. I mean, you just said my little credit union, I know what our budget is and it ain't what JP Morgan's is like. I think if we spent our entire assets under management, it's probably one months of chases, just tech teams spend in their budget, right?

So there's a little bit of a difference between us. But AI can be a really cool, great equalizer. But to your point, we have to have the data in [00:40:00] place and we've gotta have the processes in place. Like our team has to have a clear direction, vision and understanding for how can we leverage ai.

It's funny, I'll give you one really funny anecdote to that whole story. One of the reasons why I think our CEO asked me to be on the board years ago is when I represent their membership. I've been a member for, almost 25 years.

Two, I work in the industry and three, I come with kind of the technology background and experience. While I said I'm the stupidest guy in tech, I do work around it. So I learned a few things through osmosis, hopefully. But it was really funny, I took notes throughout the whole conversation.

I also had my note taker running, and at the end of it again, I ran this all through a protected environment. So my data stayed safe for it being segregated, for being a part of the credit union. But, I ran all of it through one of my Claude [00:41:00] skills and. Literally, I left the meeting and as I was leaving the meeting, I dumped all of my notes and everything in and I gave a very specific set of instructions to the skill earlier of how I wanted it to create an output for me.

And I literally just hit run. And then I got in and Uber to head home. And from my phone, the skill finished running created my output, and I sent everything to my CEO and CFO of the credit union. And on Monday they both responded and they were like, holy smokes, these are the best notes we've ever seen.

Like this. Put my notes to shame. It was hilarious. And I was like, guys, like I'm not gonna take credit. This was Claude, right? But I said, Hey, this is a great example of hopefully you could tell that I was really hyper present in that meeting, right? I was listening to everything that my fellow board members were saying that the leadership team and the credit union was saying.

I was really focused on that. I was [00:42:00] really focused on listening and then solutioning out of the listening, but it didn't mean that my notes suffered. And I was like, this is such a low hanging fruit. Simple example of how, just think about your leadership team and when the credit union is having meetings, how can you be more present?

How can you listen to solution? Just a great example of that. So sorry, I threw probably like seven different topics at you in that, but

[00:42:25] Parijat: No, but first I have to say, Josh, you're a humble person. You have enough tech skills that you leverage in your day-to-day life.

[00:42:35] Josh: I appreciate that.

[00:42:36] Parijat: The example you took is a very, very powerful example, at least from when you look at credit unions, as you rightly said. They don't have the biggest budgets, they don't have the broadest reach, they don't have a massive PR operations, brand recognition.

But I think, there's also a sense of a perception problem here, right? when you think of ai, part of AI is also [00:43:00] the generative engine optimization, or geo as we call it, and it does not work. The traditional digital marketing way, it works slightly differently. AI systems, they kind of measure volume.

They measure trust. There, there's some sort of evaluation that is done. And interestingly, a lot of these are qualities that are built in the community institutions, right? Or they've been cultivating for decades. So there are advantages, I think, that a credit union or similar has primarily when these searches come up.

So if I go ahead in a particular community and I start searching the kind of things I need, and I leverage AI to do that search, just like you used it to do your notes, but if I start doing that for my search and I give specific instructions it's natural that it'll come up beyond the PR budget because you know, some of the [00:44:00] advantages are like.

There's an authentic community authority, right? When these AI systems evaluate and they look around they would see there are community banks, and there are credit unions that work for decades, 40 years, 50 years, 60 years, right? And you have sponsors, and you have local events, and you have partners, and there are school districts all of that play in.

And so when a consumer like me, or you or somebody else is gonna ask an AI assistant for financial services in that area, the system looks and corroborates all these signals, and then automatically that local presence comes up. And so automatically that comes up as a choice. So I think what has happened over the ages, if that is digitized and that's available in the broad ecosystem of the internet. 

It is actually an advantage for AI systems to start building up when consumers start looking for it. 

[00:44:59] Josh: If you work for a credit [00:45:00] union or a community bank, GEO is absolutely essential right now. A shout out to Andrew down in the CEO of OB Credit Union, in Olympia, Washington. I sat in on a presentation that he gave in Vegas last year to a group of credit union executives and board members.

And it was really funny. He started his whole pitch with, he was like, the second I touched down here in Vegas. He's like, I literally just went to chat GPT gave it a few parameters about me and said, go find me the lowest CD rate, or sorry, the highest CD rate rather that, that I could apply for here in Las Vegas, right?

And it gave him a bunch of examples and he was like, gang, the days of winning over a slightly better rate are over. I can go to Chat GPT and everything is just at my fingertips. And it can do really intelligent analysis of the market of [00:46:00] products and solutions that are out there, the rates that are out there, what I would and would not qualify for.

And then couple that with everybody who's been doing this whole arms race of making it easier to digitally become a, you know, account holder of a financial institution or procure a product or service from them. So now it's super easy for chat GPT to be like, actually, chimes got a great rate and here, let me dump you into their account opening workflow and you can go open that account.

His point was we have to be a lot more thoughtful about GEO and we need to make sure that our unique value prop and our differentiation shines through in that. And you've gotta be different. And so the piece of advice that I would give folks is it's really this simple, okay, go pull up your chat engine of choice Chat, GPT, Gemini, Claude, whatever it may be.

Literally go in there and just say, [00:47:00] Hey, I work at this credit union. Go tell me what is your perception of us now? Where do we rank highly? Where do we fall? Who do you see as my main competitors? Who would you pitch against me? Start having a conversation. And then the real power is to go in and just say, okay, I want to influence you.

I want you to rank me higher. What do I need to do? And it'll tell you, Hey, you need to do, more whatever thought leadership on your blog that I can pull. And to your point from earlier Parijat, see you as a thought leader, right?

And elevate that up in my search results. And it'll give you a clear picture of here's where you rank today, here's what you can improve and here's how you can improve it and help you with the strategies to actually improve it, right? So use the tool to help you use the tool.

[00:47:58] Parijat: I think you're spot [00:48:00] on with that. And I think in addition to that, the other. Great comment would be at least from a credit union and a community bank perspective, is the trust that has grown over time, in the community. And all of that needs to show up, right?

Trust is a huge lever, at least in the GEO world that you laid down Josh. So I think making sure that credit unions and community banks can use the trust network. will be very important.

[00:48:35] Josh: Well, and that kind of brings us full circle to, one of the things that you started with I feel like one of your hopes is that things like technology and the next evolution of whatever our next industrial revolution is, is that it actually allows humans to just be more human and brings that more to the forefront.

You know, I probably overuse the example, but think about the advent of the [00:49:00] telephone, right? And just how incredibly connected that made us as humans. And so yes, that was technology. Yes, that was business. That brought about the eventual smartphone and the garbage of social media.

And that's a topic for another day. But it made us more connected, right? It made it so that my little brother who lives in Chattanooga, I just pick up the phone and call him and catch up, which I probably should do more of and say, how are you doing? And it made us more connected, right? Where that didn't exist before.

And so, again we've had all sorts of conversations about AI and there's absolutely pros and cons, right? And there's definitely things that I can find about it that I don't love. There's things about the industrial Revolution. I can look back and say, I don't love. Look at the pollution.

Look at what it did to other elements of just our world and our society, but look at all the good that it did too. Same thing with ai. We were talking about the energy consumption, we were talking about the land requirements for the data [00:50:00] centers and all of those things. We can also absolutely look at it as a tool that kills off creativity or kills off doing research and really thought provoking.

But it can also help with those two things too, if we use it from that lens. So I think it's an interesting place that we're in where people are trying to find ways to leverage it for the good and actually be more human rather than less human. 

I think people are still trying to figure that out, 

[00:50:28] Parijat: And the way I would like to round up your comment here is there's a concept called G One's Paradox and what it talks about is any change in technology, So when you think we had the industrial revolution there was a lot of pushback, primarily because people will lose jobs.

But what happened eventually was the job market completely expanded. The reason it expanded was what Industrial Revolution gave us was took away the [00:51:00] current work that people can start look at the next value in that position. The same happened with technology, as you rightly said. When technology came in initial pushback.

I still remember places and the cities and countries where when the first PC was shipped in they were just considered as job killers and they were burnt down. The entire warehouse was burnt down. If you see today it has provided more jobs than the Industrial Revolution can ever think of.

I think given paradox plays out even with. The future technologies that come. Right? And obviously there is a sense of fear, there is a sense of unknowing and anything that we do not know, we fear. So obviously, that's the chasm if you would, on how do we use this new technology. But, I would like to, you know, kind of close this wonderful [00:52:00] conversation that we are having, with a quote, if I may, from Scott Fitzgerald.

That's always been very close to me. It goes like this. True, intelligent is the ability to hold competing ideas in your mind and still work effectively.

[00:52:17] Josh: I'll say it once again. True intelligence is the ability to hold competing ideas in your mind and still work effectively. I believe that is the truth of today's existence.

I like that too. I think that's something that I've tried to embody more and more as I've gotten older is the idea and the concept that competing ideas make me better, right? I'm a huge fan of whatever the topic may be. I love to look at competing ideas, especially something that I'm really dead set on or passionate about.

'cause I might be missing something really obvious, right? And I love to talk to people who disagree with [00:53:00] me because man, if you can change my mind, that's pretty impressive and that's good for me. And so, look, again, we can take the pessimistic view, but it feels like so many times, especially if you're looking on social media, goodness gracious, it just feels like we're incapable of that anymore, right?

We're incapable of having competing ideas in our head. I'm so focused on the ideology of one camp or another. And we don't let those competing ideas in. And I would argue that's one of the things that makes the human race so great is diversity, is having different points of view, having different ideas, bringing 'em to the table and listening to them.

I don't know how this ties back into what role AI plays in that, but I think you can look at AI from a positive or a pessimistic sense. And I think if you talk to 10 people, [00:54:00] you can probably get a lot of different competing ideas around the role that it's gonna play or the impact that it's gonna have.

Yeah, there's gonna be a lot of negative, don't get me wrong, but I'd like to focus on the positive. 

[00:54:11] Parijat: Yeah, and to your comment, I think it is good to have competing ideas. Only because you do not have a single form to align and adhere with to some extent. Every new technology needs to be challenged. And that's what's happening right now in its way in form. But I'll bring it back to where we started, right?

It goes back to what are the solutions that we are trying to drive for humanity? What are the solutions that we are trying to drive for communities? What are the solutions that we are trying to drive for our families, right? So it's the, the family is the microcosm, the country and the world is the macrocosm, right?

But as you keep building solutions, it should affect you and your [00:55:00] family and it should eventually affect the community, the society, the state, the country, the world in its larger form. And if we can get this trajectory. I think fear is good because it keeps us on the right guardrails and makes us ask the right questions.

But I am equally optimistic in the usage of this technology because this has so much to offer and so much for humans to even get into the next value-based activity, if you would. Where you can, again, going back to where we started, look at you on the face and listen to you.

I can give you my total attention. I can make sure that the best gift I give you is my attention and vice versa. Yeah, I am pretty optimistic where this might land.

[00:55:45] Josh: Yeah. I like that approach. And I think that was one of the things that really drew me to the conversation with you was just that concept, right? Of how can we leverage technology and everything that's come before it and is gonna come after it to be better [00:56:00] humans. And I know that sounds really like pie in the sky, ethereal, but I think that's a good guiding light and I would be remiss if I didn't come back to one thing that we touched on earlier before I let you go though.

You said you get to see kind of a global view of this. what you find fascinating is that while we all have different diverse lenses that we're looking at it through, some of the challenges and opportunities are common. Any kind of parting thoughts on that topic?

[00:56:24] Parijat: I think, going back to my comment on looking at any problem statement through a process and a data door, when you start looking at that, the challenges are very, very common across the board, right? When you lay out the process, for example, the issue of fraud that we see in us is equally valent in LatAm is also very similar when I do the discussion at Amsterdam, right?

The data rules are different. There might be separate rules that Europe uses versus US versus LA and areas like that or Asia. [00:57:00] But the problem statement is very similar, right? So the solutions that we build, if we build it with slightly broader scope, can be applicable across the board.

Right, and we can help solve those problems. Some are very narrow and specific in that you may not be Europe. As you understand, every country has its own rules and norms and structures, so it plays very differently than it would in us, or Asia Pacific. But I think the problem statements are very common.

More importantly, because we are focused a bit on financial services, financial services being the pillar of society, the needs and the ask from people across the world are very similar. I should be able to access, I should be able to get access and use financial services.

I should be able to get money and I should be able to drive something for myself and the community. [00:58:00] That I'm trying to build other country at large. When you start solving the problems across the board. The problems will be specific. It's gonna be a very different problem, what you wanna solve in Bangalore versus what you want solve in Amsterdam versus what you want to solve in Mexico City.

Only because the societal structure is slightly different, but the solution of it and the final impact of it is pretty much the same.

[00:58:24] Josh: May sound silly, but you can boil it down to financial services. Money is a big part of our society no matter where you live. And being able to access it when, where, why, how you want, and do what you want with it, when you want it.

It's kinda the crux of it all right. So we're all trying to solve that problem and offer a value add within that stream. Parijat, before I let you go, I got two final questions for you. So one, where do you go for information? How do you stay up to date on what's happening in the world?

[00:58:55] Parijat: That's a difficult question because I obviously go through [00:59:00] multiple reads and write up and blog posts and podcasts without naming anything specific. But I think my biggest information comes from practitioners. I spend, more than two thirds of the year talking to people across the board, people like you, people who are industry practitioners, people who are in the system, working the system, finding the issues, and they're cribbing about it, right? I think that is where my information comes in, and it's a wonderful spectrum. I have people who are amazingly optimistic with the solution that they have, and interestingly, that same solution is completely shut down on the other part of the world. But that is the interesting part, right? Because that's where I get to feel and understand the value add that moves across in different parts. So, I think that is very interesting to me when I keep talking to people across the board, coming [01:00:00] back to, again, the factor of listening. I love to listen.

I would like to hear what your problem statement is, and I would like to then figure out the solution and structure for it. That's been my core. Mantra.

[01:00:13] Josh: I mean, I feel like I should have known what your answer would be given this whole episode and that your answer would be just talking to people, so that's funny. Last but not least, if people want to connect with you, if they wanna learn more about your organization and what you're doing, how can they connect with you and learn more?

[01:00:30] Parijat: So the best way to connect with me is on LinkedIn. That's probably the only social media platform that I'm on, other than the Forbes platform. And the company I'm part of is LatentView Analytics, which is a pr play Data first organization 20 years, more than two decades in that area.

And we work across four big verticals. Anybody wants to know more about LatentView? It's the website. That should be good. Other than LinkedIn and other social media, we are [01:01:00] pretty much in every other social media structures, that'll be a good place for anybody who wants to learn more about LatentView Analytics as an organization.

[01:01:08] Josh: Parijat, it's been absolutely a pleasure chatting with you. Thank you for providing your unique perspective and ideas and thanks for all the different kind of analogies and callbacks to different things people have said and done. I thought that was pretty cool.

But thank you again, just so much for coming and being a guest on the Digital Banking Podcast and sharing with us today.

[01:01:28] Parijat: Thank you, Josh. It was wonderful to have this discussion with you. Loved it. It was warm, and hope to come back again. Take care.

[01:01:37] Josh: Thank you, sir. 

[01:01:38] Speaker 3: 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. 

 [01:02:00]