Episode Transcript
[00:00:00] Speaker A: Donna, thank you so much for joining me today. This is going to be so much fun. I want to start off with a difficult question. What is the most difficult customer and moment and tell us that story. How do you kind of work through that? And it ultimately was solved.
[00:00:55] Speaker B: So they have system. If we need to fully rebuild from scratch and build it on the microservices, say that in terms of the roadmaps and just clarification on which functionality could go in the alpha version or MVP or beta version, that's not an easy story. I would say building projects and products from scratch on the new company or in the new startup is way easier from my point of view.
[00:01:26] Speaker A: So where's the moment of conflict where the customer wants to do A and you really believe that B is the solution? What does that look like? Where does that.
[00:01:40] Speaker B: You know, the hardest thing is when as far as they already had some application and some web bank, they have already some user experience. Experience. It's very outdated, it's like 10 years or 15 years old. And when you're trying to promote them some kind of new solutions, new user experience, it's kind of hard because you know, they have some kind of understanding of how the product should work. They have already some experience and they do not understand what for we need to change. But at the same time, when you're asking them what is your like main customer, how old is he? They're saying like 18 to 35 is the main customers who we want to acquire right now. And then you're asking them, okay, then we need to provide them new solutions, a new user experience. Just check out the applications which they're using in everyday life. It's YouTube, Instagram, meta and different kind of apps in the iPhone.
And you're not providing such kind of user experience which are provided by those apps. So here is the question how to be new and nice and at the same time on your user experience. And the answer is there is no way.
[00:02:56] Speaker A: So let me ask you, let me ask you this. Is the problem clear? Like do they have like surveys saying or feedback from the customers or is it not actually clear that there is a problem? Let me start there.
[00:03:11] Speaker B: I would say that customers are replying that they want to have a new application, a new web with a new user experience in the new interface. And at the same time when they're checking the competitors and the local market competitors have almost same application which they have right now. So from the one point of view, I can agree with them in terms of having an old application, but 100 features, it's a good thing. But on the other hand, maybe next year there will appear a new fintech or new NEO bank who will implement all the, all those features and in scope of the new user experience and new user interface. And I can say that there is a huge risk that some percentage of the customers will just go to this new fintech.
[00:04:03] Speaker A: That's, that's absolutely beautiful. I mean, you know the, the Innovator's dilemma is a very old book, but it's crazy how we keep seeing that same thing playing out. And everybody knows about first mover advantage, but I think what people don't think about is the fear of being the first mover and making that mistake since nobody else has done it yet.
I think that's beautiful. So you're a big organization, everybody else has got kind of the old tech experience that you do and making that first move is scary.
How do you approach that challenge? How do you kind of work through making these, I would say somewhat brave decisions that could also be dangerous for the business.
[00:04:48] Speaker B: I can explain you. Actually the hardest thing is to be the first one in the world. When you're already like second or third or like in the top 100, then it's a bit easier because you have competitors on the worldwide market. You can check their user experience, user interface and you can compare and you can provide different kind of analysis and based on that you can let's say get an approvals and defend your user experience and your innovative approach. So that's what we are doing. But that's not all the time helping because as far as I said this bank is in Armenia and the market is a bit different comparing to Europe, UK or US market.
[00:05:30] Speaker A: So what are, what are, how is the, how are you and the company perceiving cultural differences? That's so interesting. How does that come into decision making?
[00:05:43] Speaker B: I would say that that's a bit different because I was working previously in the Scandinavian bank and I have a huge experience in Europe and understanding how to get an approvals and how to build products.
There is a bit another approach.
You should be 100% clear that what you are providing them and what you're depending is accurate and you really know what you are doing because if there is 1% that they can ask you some questions why this product is working, not in a way as it's working right now and you don't have answer on that, then with the high possibility you will not get enough proof.
So I would say that people there are, let's Say valuing when you know what you're doing and you should have like huge, let's say, points on what you're doing.
[00:06:46] Speaker A: I mean, if I thought about, you know, what kind of people I would run to, you know, running my, managing my money, I definitely want, you know, the people to be very accurate and get their shit together. So it definitely doesn't sound like a bad thing to have a demanding, let's say, kind of customer when it comes to banking or even healthcare. Right. Because there are high stakes in danger.
I wanted to change the topic a little bit and ask you about security.
We're hearing about hacks happening almost on a weekly basis, definitely on a monthly basis. How do you kind of think through creating a product that is true to secure? What are the elements that go into it?
[00:07:34] Speaker B: That's also a big topic. Interesting one because I was studying cybersecurity and I'm a fan of this topic.
I would say that we're using the latest technologies on the market, but at the same time those technologies are, let's say enterprise certified, meaning that you cannot use some kind of library from GitHub while building a product for the bank. So you should be also very accurate about your technological stack when you're building something right now.
So first of all, all the best practices and approaches which used all should be all the libs, all the, and let's say software solutions which we're using, those are up to date. We're constantly controlling, we are constantly checking all the new updates on the market, visiting conferences in scope of development just to know what our colleagues are doing all over the world. And we have security specialists who are also making different kind of penetration tests and load tests to understand are we secure, can we work with the big loads of customers and data? And that's a big topic for the banks is if we're speaking about some small startup which is dealing even with the E commerce, they don't have such kind of problems and certification issues as banks would have if someone will hack it.
[00:09:11] Speaker A: In fact, the thing that I hear in a lot of startups is that they, you know, they just let us get to the fact that we have that problem that we have so many users that it breaks and then we'll fix it. So they're kind of thinking about it from a growth mindset as opposed to, let's say a security and stress load mindset first.
Where do you start? Go ahead.
[00:09:36] Speaker B: I will just add the price of the mistake in the banks and the fintechs is way bigger in Terms of the fines in Europe especially.
So yeah, I think that I know that majority of the banks in the central Europe pay a lot of attention to the security.
[00:09:54] Speaker A: Yeah, it's really interesting because there's this on the one hand, like everybody wants their data to be protected, but on the other hand having more and more regulation, what it's actually doing is acting as a barrier to entrance, preventing innovation. So it's actually giving more power to the incumbents and preventing innovation from that perspective. And I say that with my hand on the heart knowing that I don't have a good solution because we want both of those things. We want competition, but we also want security.
What about compliance? When you think about compliance in the market and different standards, SOC ISO 2701 GDPR is different because it's not about security. Do you think the compliance is really making companies more secure or is it just checking the boxes?
[00:10:50] Speaker B: You know, that's a painful topic for me and I think for all the product managers and chief products and all the management stuff for dealing with the new product development.
Because from one point of view, I fully agree that those standards should be on the table especially for the banks, fintechs and the healthcare and insurances.
On the other hand, all of those standards should really expand and save customers data and customers from frauds and from cyber attacks. What do I say if we're speaking about the data privacy that the customers should sign in the contract like 10 times that they agreed to provide data to the bank? That seems weird for me. I mean it's better to make a good UI and good UX and provide customer a noticeable field that you are giving us a possibility to work with your data and we are saving your data properly and we hope that nothing will, let's say, happen with your data.
From my point of view, for a customer that's the most important because nowadays wherever you register yourself, you are providing all the data. The majority of clients are just ticking the box and that's it. But what is standing like on the back side of this box in the 90% of the cases is just like some let's say CMS with the outdated CMS with zero day, let's say vulnerability.
So that's my point of view on compliance and how to work.
[00:12:44] Speaker A: Yeah, I mean it's, it's really a challenge because on the one hand we want to have these laws be so descriptive about getting, you know, permission and signing all these things and explaining. Exactly. But your average customer, even when it's about your bank account, they don't read it.
Yeah, and you know, they're not going to read it. They're not going to understand the very legal language. And it feels like we're just missing something in this process which is not really helping us to move forward. And really I'll be a little cynical and I'll say in some way we're just opening the door for more, you know, work by lawyers, you know, for people to sue each other than actually creating real value and security.
[00:13:33] Speaker B: You know, I just came up with the idea that in the year of the ChatGPT will be a nice function when you have some kind of big document you need to sign to have a button summarized for me and to get like one, two sentences about what's written there. And that will be very useful for customers, I think.
[00:13:54] Speaker A: You know, I like the obvious. The obvious problem with ChatGPT, and I'll say very explicitly right now is this inability to go and kind of fact check. But there's other software that is coming up where it will actually tell you, oh, this piece of information that I've given you comes from this document. So you can kind of click on it and then you can look, oh, does this actually make sense? I think that's the future, right?
[00:14:18] Speaker B: Because if you remember, that's called explained AI, when you can correctly check from which source this data is coming. And that's wonderful. I think when that will appear, a lot of banks in the fintechs and all the legal companies who work with the legal stuff, they will start using that.
[00:14:38] Speaker A: There's something delightful about it. Because, you know, if you're expecting to get a 100% right answer from an AI, then, you know, we don't have that right now. And I would argue we don't have that with Google search because you're just going to some buddy's website and maybe they're wrong. Right? And we don't have that with human beings. Like, you can't ask a fellow human being something and they'll. So, you know, in some way our expectation from AI is maybe too high. And I heard this guy, he was saying, look, if in 200 million km a automatic driving car has an accident, like everybody's like, oh my God, the robot had an accident. But my point is that if you had humans driving 200 million kilometers, how many accidents would the human have? And if it's 10 times better, then it should be clear to everyone that the computer is doing a better job than a hero. And it shouldn't be as emotionally shocking to us, but that's not how psychology works. So let me ask you this following question. When you think about as a product manager, integrating artificial intelligence into processes that are highly sensitive like banking, what's the thought process here? And also what are the rejects you get from your, from your team members and clients?
[00:16:03] Speaker B: I would say that we are not using for clients AI currently.
I mean like for clients, products, features, but we are using some kind of AI in the back office and back inside.
Why? Because as we're mostly working in the Europe that's very highly regulated market and we still have like documents which are saying that law even which is saying that you cannot make any decisions with the robots. Like all the decisions should be made by human being.
So to the time when we have explained AI, we are just not able to do that. But if we're speaking about the programming and how we're developing different kind of products, and even about the analytics, we're highly using AI because that helps us a lot in the competitor analysis, that helps us a lot in the programming. Because when you can provide like a huge part of your code to your AI partner and he can then provide you some hints while you're writing all the other new code, like comments from our developers, is that the efficiency grow on 20, 30%.
And that's not because AI is writing code by themselves. No, like AI is okay, you can ask it for writing some part of code, but result could be so. So the best point here is that AI can realize and understand all your code and provide some hints and you from your point of view as a senior developer, already can accept it or revoke it, or accept it and make changes from yourself also. And then you will get the best result.
So still like decision making is on the humans, not AI.
[00:18:14] Speaker A: Currently I was quite amazed I actually asked it to write code. And I'm a software developer in my past, so if I see code I can read it, understand it fully. But what was interesting is you can kind of tell it. Here is an API that I need to connect to and it's got 50 pages of documents and you want something very simple. You want to say, oh, go get, you know, go get this information. It can go read the whole API, it can pull out what is the exact correct function that you need to evoke, give you some sample code and then you can kind of look at it and say oh yeah, that's right. And it just saved you, you know, reading 50 pages of documentation as opposed to, you know, gave you a suggestion. You're like, oh yeah, I can use that. So for me it was just mind boggling. Like that process, how, how it can turn it from like a two, three hour process to literally a 30 minute process was quite incredible.
[00:19:07] Speaker B: I can even add that we are using such kind of approaches. So for example we had a topic with core migration and one of our partners, they had very old core ledger system and they needed to make changes and to swap an old one to the new one. How to swap if you don't have enough documentation on your old core what.
So what we have done, we have written our own AI model where we just fed our partner salt poor and understood all the topography and architecture of it and it also automatically generates the documentation so we save lots of time and lots of engineering time.
[00:20:00] Speaker A: What was the documentation created based upon? The code itself or something else?
[00:20:05] Speaker B: Code itself and databases on which kind of connections there is inside of the core and how the data from one for example table is moving to another table.
[00:20:17] Speaker A: It's just mind boggling. I have one final question to you.
When you think about the young generation coming into the professional work environment, what is the one piece of advice that you would give them?
[00:20:37] Speaker B: You know I can just from my point of view give such an insight that if you want to make an innovations in the bank first of all should understand how the bank is working itself and I will suggest all those people to go first to the bank or their at least one year, two years to understand all the bureaucracy, all the processes and all the stuff which is inside and afterwards you will be able to work on some startup or with some kind of, let's say new NEO bank because you will understand which processes are working better in the old bank and how to provide these good processes to the NEO bank where you have all the new stuff, all the fresh new generation developers with new stack and with the possibility to build the new amazing features and products.
So yeah, that's my approach to those people and don't be frustrated, be prepared that building products in the fintechs could take huge amount of time but then people who will be using those features will like you and they will like your product you're building.
[00:21:57] Speaker A: What an absolute pleasure to have you today. Thank you so much for joining.
[00:22:01] Speaker B: Thank you so much.