Afraid You're Falling Behind in AI? Watch This First (with Viktoria Otstavnaya)

Speaker A: Your project was one of the most compelling in terms of something very practical to pretty much everyone can recognize this problem because I remember being like, oh, uh, no one has not had this problem before.

Speaker B: You see what you receive in terms of idea, how you communicate it into what is being created and then you can just literally enhance it. Enhance it, enhance it. I think we live in the world where there is so much information going on, entering our brain and a lot of things that we just have to face that it's important to understand what is important and what is not important.

Speaker A: Welcome, um, to another episode of AI Product Leader. Every week we shine a light on what it really takes to join the AI world. Real conversations with AI builders, from founders to product leaders, AI experts, technologists, and everything in between to understand where the world of AI is right now and what that path looks like, whether you are building AI products or a career in AI. I'm Polly Allen. I'm the founder of AI Career Boost where I spend my time helping product leaders land the roles of their dreams in AI and thrive leading AI products and teams. And I spent my career on both the product management and software development sides of the fence. I'm an ex Alexa AI principal product manager where I successfully launched the very first generative AI answers on Alexa back in 2020. And I'm very excited to have our guest joining us today. Victoria Otsonaya is a principal product manager who specializes in modernizing complex systems and delivering high impact product transformations at scale at consumer cellular. She's led major initiatives that reduce product launch times from months to minutes, accelerated SIM activation from seven days to one, and helped drive large scale consumer migrations and national retail growth. Known for combining operational rigor with strong cross functional leadership, Victoria has built and scaled products across telecom, fintech, healthcare and utilities. And she's especially skilled at bringing clarity, structure and momentum to fast moving environments where execution really matters. Victoria, thanks so much for joining us today. I've been looking forward to our conversation, thankfully.

Speaker B: Yeah, likewise. I'm so excited to be here.

Speaker A: Oh wonderful. Well, I'd love to kick off just having our listeners get to know a little bit more about your background. I think it's so interesting you, you came from Belarus originally, is that, that correct?

Speaker B: Yeah, that's correct. I was actually born in Germany and then um, lived there for the whole three months and then my parents back to Belarus. Yeah, so I spent uh, all my childhood you um, um, student years over there. I uh, graduated from the Belarusian State University of Informatics and Radio Electronics, worked at EPAM Systems, which is a uh, global consulting and software development company. Uh, and I kind of started my uh, career uh right uh into the software uh development world, uh right away and uh, stayed there and planning to stay there.

Speaker A: What drew you to the software and technology space in the first place?

Speaker B: So uh, initially uh, even before uh EPAM Systems I was uh, working as uh Android developer. I also uh, did some software testing but my dream was always uh to be in the business analyst, uh, product owner and then product management role. And when I realized that this is where my passion actually is so that when I received a ticket as a developer it's like why are we doing that? We could do this, uh, we could do that instead. Uh, and so then when I realized that I love to not only implement stuff and think of it but also discuss it, uh, and what I love about uh, product management role specifically is that there is a combination. It's like you know, 50, 60% of your time you are actually thinking of something and living in the future, uh, of creating new products and discussing how they will work and another like uh, 40 to 50%, uh, you actually thinking okay, how all of those ideas work, uh together. Uh, and so and it's wonderful because you like at least for me, I close my natural desire for uh, like talking and discussing and collaborating but also like natural desire just to be like you know, on my own thinking about the things. That's where I feel super grateful as well. Like being able to do that role and work in it.

Speaker A: Amazing. Do you feel like that depth in technology, having been a developer, helps you as a product manager?

Speaker B: Absolutely. So I think that um, uh it is at least for me I uh, like to understand uh, the uh, way how things work uh to the level of uh, like hey that's our platform, uh, that's how different services talk to one another. And that gives me understanding uh, when I'm talking about uh, you know, all possibilities and keeping the general ideas that we want to have, uh, but also knowing the limitations, uh, and kind of quickly being able to tell uh, what is possible, what is uh, not possible, uh, for the short term, mid term and long term, uh, because without that knowledge it is very hard. It's like too much of uh okay, let me talk to the team. Then you go and like no, that's not possible. But what if they think like changes and when you know all of that, uh, and you can assess it because if you used to do that uh, then you would. Oh yeah, then it's like it's, it can Be difficult. So. And you can right away say it and think uh, of it like by yourself first and then when talking to uh, anybody from your stakeholders.

Speaker A: Yeah, absolutely. I feel like there's also an element of like developer empathy. You have. You're just like oh, can I ask them to do this? Oh they're not going to like it. How do I break. Have to go and rework this thing that I know they're very proud of. Absolutely. So with a technical background you'd come

Speaker B: into product professional empathy. M. Exactly.

Speaker A: Empathy. I love it. So with a technical background you'd come into product management. Um, and so I'd love to know when you first started, when AI really first create crossed your radar. Um, because it was sooner than a lot of PMs today who are just becoming aware of AI starting to incorporate it.

Speaker B: Yeah, I remember that. So it was in 20, uh, 19 so way, uh before the overall hype about AI happened like years before that. That was actually uh, I think it was. Yeah it was the beginning of the 2019. Myself and my boss at the time, the CTO of uh, the company where we used to work together. Uh, so we were thinking about how to automate uh the whole foreclosure processes uh, in uh mortgage servicing. So which requires a ton of documentation, uh, like reading uh, and understanding which processes are uh, applied in each state. Um, uh and people have to go through all of different communications that is received, create a case. Uh, why wouldn't we just automatically do all of those things knowing different jurisdictions in every state? I think uh, our CTO found the company and on our actually it was Disrupt conference that was happening in San Francisco. Uh and we uh, scanned through uh the vendors that were over there, found some of the uh, AI and machine learning uh vendors uh, that can help us with that. Uh, and I think out of probably two or three at that time, only one was uh, that could really automate all of those things. So we met with them and started working on uh, how we can read tons of uh, pages of different PDFs, scanned documents, documents in PDF format. Just literally uh, see uh, like what are the laws, uh what is the case uh, for each uh, like you know, um, mortgagee, like a person who is actually owning the mortgage automatically create a case, uh and uh, see if they basically have to ah, enter the foreclosure process and which documents we need to send to them. And that's how it started. We uh, implemented that and at that point of time I thought like that is really cool. But I did not see anybody actually uh, doing something like that before. So. And when the whole AI hype started, um, I did not even uh, think that this will uh, be something huge because like knowing of what happened in 2019 and 2020, we did something and I thought like oh, it will be like another um, couple of companies will introduce something like that. But this is where it will stop. But that's not true.

Speaker A: I think uh, I'm similar where if you're already grounded in the industry, a lot of stuff that was coming out, especially early Chat GPT you were like huh, huh. Yeah like so yeah, yeah, it's like Google. I.

Speaker B: That's why I did not understand people like oh yeah, I'm using Chat GPT. I'm like yeah, okay so that's like uh, like not that special.

Speaker A: Yeah. A certification or. Exactly. Yeah. That it's um. But yeah even the fact too that at the, at that time reading a PDF was like. And uh, you don't have Automatic reading of PDFs and now.

Speaker B: Yeah. Like asking questions, you know and yeah, things like that.

Speaker A: Exactly. Or fill out a form based on the contents of a PDF, et cetera. Yeah, ah absolutely.

Speaker B: Well fantastic.

Speaker A: Now you've done more work with AI since then, right. I'd love to hear some about some of the other initiatives you've been working on. Yeah.

Speaker B: So at um. I won't be able to go into too many like details uh from uh my workplace. But uh some of the things that uh we uh integrated with uh one of the Equifax products to help um fraud uh during the web checkout on our consumer cellular side uh and uh, that was uh also great to work and understand like that. Hey, what are some components that vendors who are LLM based uh solutions what they need uh for their assessment modules to work better. That's good. Like though even if you do not own the actual custom mod, uh you understand what the model needs to perform better. And it's even when you are reading through the documentation, uh what is mandatory, what is optional and talking to them uh, of which rules they have, what they recommend kind of gives you clues uh, uh to uh, what they're using, uh overall uh for maybe how they're prompting and uh. So that's important and interesting. Um, uh so even if uh, uh and that's one of my advice if like somebody, if your company uh, or your personal project, you are not creating your custom model and you're just integrating, uh it's like to take an extra step and understand okay, based on what the APIs expose what uh, other things that they're asking you to uh, deploy specifically third party library. Uh and you understand what information they need, what they derive from that. Talking to all people who actually configure platform, uh, for you is what they, they what their rules and um, asking a question, okay, what your other clients use, uh, and what gives the best result that will help to understand that what can be used. Uh, and like how can you also think when you are creating your own um, AI based product, uh, what is the most important and how you can use it?

Speaker A: Absolutely, absolutely. Seeing that that sounds like that industry knowledge. Ah, like actually understanding the user context is, is still the important piece as product. Right. It hasn't been so technically focused. It's still the human gas.

Speaker B: Yeah, exactly. Like you know people who are configuring like you can only like literally sometimes mandatory fields will be like two or three or even one. And so uh, but whenever you start configuring like to actually use it. This is where all of the gold comes in. Ah, uh, okay, so most likely you still use this uh, and this is your uh, marker like you know, factor because nobody will disclose you what's like you are behind the custom model but by talking through that you can understand uh, what things they're using anyway, uh, and at least what is a defining factor.

Speaker A: Oh, interesting. Um, and so you've also been using AI outside of work. Can you talk a little bit about what motivated you to do some of your own AI projects and things outside of work?

Speaker B: Absolutely. So uh, first of all, uh, uh, whenever this whole hype is uh, happening and I think everybody goes through these phases uh, of okay, so it's there, um, then you start using it and you use like you know, use it as junior user. And then you think of uh, okay, what else can be done? And you see on LinkedIn it's like every single post I saw on LinkedIn for a long time was AI AI, AI and was like oh my gosh, like do people talk about anything except, except that. So at least for me and uh, I talked with a lot of people around that um, ah, there is a lot of like you know, fear based motivation uh, that is happening and like it might not be fear, it might be like oh I'm blind, like slight discomfort or um, like any other feeling. But I have not yet heard anybody telling like you know, positive emotion, uh, behind uh, behind it. Yeah, like some there is excitement whenever you start working on that. But a lot of the things is like oh it's something new. And it's natural for, like, for us for people to start, like, from uncomfortable feelings. The most important that you go there. And so that's where for me it's like, like, I need to do something because I don't like that I'm feeling behind, uh, the whole AI train, uh, and I need to jump on it, uh, and. Or talk or take another train. So the first one already departed.

Speaker A: This is very exciting.

Speaker B: Yeah. So, yeah. So that's where I, uh, figured that, like, I want to get some knowledge. It's interesting. And I was looking through, like, you know, reading some articles and it's like, oh, it's actually interesting what people are doing then moment I saw, like, you know, some of The, I think LinkedIn, uh, ad, uh, about your course. Yeah. So it's like, oh, it's actually interesting, you know, when I just had a thought that I want to do something, uh, and like, started reading and like, read more on this article. Uh, that's the ad started to show up. And for me, I think it was just overall, uh, motivation that came out of some, like, discomfort of like, not having the knowledge, uh, of it and understanding that what drove me to actually look for some courses to, uh, dive deeper. And I looked up like a ton of them. Some were just very short. Uh, some were also, like, rather deep. But what I liked about yours, that it was practical. And for me, like, when I touch it with my, like, you know, with my fingers, with my hands and like, I do something not just like, you know, consume theories, which is good. Like, you know, to have knowledge, uh, is important. Uh, of course foundation is important, but it's also like, okay, what can I do? Uh, is that, you know, you hear all of those things that, um. Like, I have never coded. And I can do stuff. It's like m. I doubt it. And so, uh, and. But like, knowing how to create stuff, it was also like, let me see, where is it? Like, is it really true? And it just motivated me, uh, as gaining knowledge always motivated me, uh, or either of the things that like, really, really interest me or things that I feel like I do not have enough knowledge. And it is important to have that knowledge in our life. So. And with AI, it was a second. Uh, so it's, uh. I, like, I thought that, well, it's. It's time. So even though it's, uh, a little bit scary and it's like, discomforting, but I want to learn and uh, practical, like, application of it was important. So.

Speaker A: Amazing. Awesome. Yeah. So you joined our Blueprint cohort. Um, I know one of your project was um, one of the most like compelling in terms of something very practical to pretty much everyone can recognize this problem. Can you talk a little bit about your project? Because I remember being like, oh, no one has not had this problem before.

Speaker B: Yeah, sure. So at that time, uh, I thought of. To be honest with you, I could not find um, an idea very quickly. Uh, and this idea that was a project that I came up with afterwards, uh, came to me like right away. But I was like, maybe it's like it's not too deep. Um, and for me it's like it should correspond to my values, blah, blah, blah and etc. Uh, so I was like, okay, this or that. And then everything else was just like, yeah. I was not excited about the idea. So I decided I'll go with the first one which is uh, Google Drive Optimizer. So I uh, had massive Google Drives as maybe a lot of people are though. I like everything to be in order and I try to maintain it, but whenever you let it go for just three months, it becomes a mess again. Uh, and I thought, okay, what can I do? And I decided is that I'll go with that. Even if it's not really like, you know, excites me and that it corresponds to my values. It will be a good thing that I will still address a problem of me and on other many people I will practice it. And that's good, uh, enough for that. So once I get knowledge, I can also like think. And when my creativity is uh, in the boost mode again, I can come up with something uh, exciting and I will have some knowledge and foundation. So yeah, and that's how uh, I ended up go like, you know, selecting this idea and uh, implementing it. And that was really cool of uh, proving to m. My skeptical self. Yes, people don't really have to have any coding knowledge. And though I had it as like I literally could have just entered it, uh, without having any uh, coding knowledge whatsoever. The most important is that you see what you receive. Uh, it's literally a busy week of uh, uh, like what you see is what you get, uh, in terms of idea, how you communicate it into what is being created and then you can just literally like enhance it. Enhance it, enhance it. And I remember that um, once, the very first time that it finally ran, I was like, wow, that's cool. So the main pain point was not even about how the whole system works. It's about how to connect, uh, through Google OS or auth. Ah, Service set up authentication. I was like, why? It's not working in test environment. It's only like an. I, like, literally was only tested in prod because it was not possible to set it up, at least for me

Speaker A: to set multiple environments up. Exactly. It's also new. That's what's crazy. You realize when you come from a tech background, you're like, well, where's all the infrastructure to do what I want to do with real software? Oh, we're not there yet. We're still experimenting. Right. But yeah, the Google Drive optimizer. I loved how you were thinking about this notion of organizing your Google Drive and then we got to do a deep dive on like, how would you evaluate this and like, um, edge cases and things like that. I still use this as a, As a teaching example in the course because people sometimes think, like, when you say, you know, test data, I'm like, can you think of like, all the different ways people might organize their Google Drives or like, have them configured that might be disorganized or organized? And then they're like, oh, wow, okay, I get it. This is a big part of the problem is figuring out the test data, things like that.

Speaker B: Uh-huh. Yeah. Yeah. And sometimes AI can actually help with that. Uh, so one of the things was for me, it's like, where will I get all of the uh, like, you know, test data for like, you know, for evaluations? Because I don't have time to go through all of those documents or even upload them, uh, to like, you know, chatgpt at this point of time for each. To understand what it is, because then I have to open it up to make sure that it's uh, like what it. What this document is and put it as like, classify it in this way. So one of the things was, ah, like, you know, put it in ChatGPT and ask generate me 100 PDF, like names. Uh, and uh, some uh, the documents with uh, their text, like what they are for. You don't even have to have, uh, actual documents, uh, um, for this type of uh, evaluation. So.

Speaker A: Yeah, exactly. In synthetic data can get a long way. Um. Very cool. So you've pivoted now. You're not going to do that anymore. I know. Since Cowork and Claude were released since then, and that's one of their use cases now is like, clean up your drive. Clean up your Google Drive. Um, but to a new. A new idea. Um, I know you can't tell us much. It's under stealth. Is that right?

Speaker B: Yeah. Ah, so it's uh, something that um, myself and my husband, we are doing together. Uh, and uh, we have uh, like spent a lot of hours generating different ideas, chatting with um, he chatgpt me m Claude and like talking about what our strengths are, uh, what our beliefs are, uh, what are our values are, uh, and what we want to do and kind of like doing the deep uh, research of um, our ideas and how we can probably combine them all. And so we found one thing that uh, excites us both. Uh, and we are currently uh, working on it. Uh, so we are at the uh, prototype stage when they are all generated. Uh, and we right now going to work on actually putting it to life. Uh, so we'll see how it goes. But, but that on the. Yeah. And on the contrary with Google Drive, uh, Optimizer, that was really good example of where uh, you can just practice skills to understand how everything works and like literally just practice it. Uh, and then when you have uh, your creativity, when you are not tired, you can. Okay, now I'm ready uh, to like, you know, to generate some ideas and combine like my knowledge with my values. Yeah. So that's the thing that is important. Like you can start with one thing, uh, and then uh, once you get knowledge and uh, important tools, uh, that you can use, anything is possible.

Speaker A: Anything is possible. That's super exciting. Well, it's very exciting to see your growth from coming out of. Hey, I have an awareness of AI, but what's this hype about Gen AI now your own genetics, possibly a business, a side business, um, as well as doing working with AI as your work. How, how was this change possible? What, what, what would you advise to people if they want to uh, build their confidence in AI? Get hands on, get with it. What, what would you tell them to do?

Speaker B: Yeah, I think first, uh, thing is that to understand what is the best way for them to get the needed education. Uh, so for some people it is uh, mastering it on their own and just reading a ton of literature. Um, there is a ton of stuff when you Google it, when you like, we can even ask Claudo like any a me a plan, uh, based on uh, like write me a plan of my education based on how I want to um, get the knowledge and how I literally like, you know, uh, thrive in which situation and how many hours I have. So basically just like know yourself, know how you um, can get this knowledge. If you are a person who like, you know, wants to study, uh, like with somebody who can answer questions and definitely take a course, um, and uh, so Polis one is amazing. So it's like in no way advertisement or anything. So it's just um, she's amazing and she creates like a really good um, collaborative uh, culture. And it's uh, very easy uh, to just um, like learn stuff or like, you know, any others that you feel comfortable about, uh, and just do it like not to be afraid. Uh, so it's really not difficult. I would also say like, um, just ask yourself, like, is it better to do by myself, uh, or is it like something that can be automated? Uh, and one of the things in my head, uh, that uh, I also um, listened to a podcast, uh, about what is better kind of older generation, not old who use Google Search versus new one who uses different AI tools. Ah, is that uh, there is definitely. They think that um, like we can forget how we do how we did the stuff. Um, uh, but that knowledge, uh, is still there, but with time it can not be as sharp. Uh, so I think that for me, um, and the way how I use it right now is that I always think by myself, uh, when it's really important and I want to have my personal opinion about something that I will create bullet point list for myself, um, of the things that when I analyze, uh, things or you know, do an assessment on vendors or on like any features and then only then I will ask AI okay, like challenge me. What am I missing? Uh, what other things, like what alternatives, uh, you can suggest. And for everything else, like what does not really matter for you, you can use AI. Uh, that's like, you know, all of the noise can just go away. And for something that does matter, like still, uh, maintain uh, the skills that you have, keep it sharp because was uh, I don't know what is. It will be like huge disruption and like you cannot use clot and chat GPT for three days. Like you know, you like you will still have your brain.

Speaker A: So has that happened to you yet? It's definitely happened to me where all of a sudden you're like, I need to cloud this. I don't want to do it out of my own brain. Yeah, yeah.

Speaker B: I'm like afraid of that time where like we will be like so much dependent. I don't know what I would do if there was no Google Maps. Uh, so. And uh, like that like terrifies me because I'm so bad, like with all directions. So I need my Google Maps and I'd like never want to rely on like ChatGPT or quote that much.

Speaker A: So I think that's really Good advice. So if I'm summarizing the advice, it is like, hey, if this is like something that really matters, that it's high quality, but also something that it matters that it comes from you, whether that's like the accountability of it or like, like I don't know, like delegating, buying your best friend a birthday present, like, kind doesn't feel like it's from you. Right. So there's still things that we should focus on like hey, let that come from your brain first and use AI's collaboration partner. But then it's like hey, this is the 80% of things that just need to get done. Right?

Speaker B: Yeah. Because I think we live in the world where there is so much information going on and like uh, you know, entering our brain and a lot of things that we just have to face that uh, it's important to understand, um, what is important and what is not important. And like to summarize is like, yeah, first think know yourself and know how you learn better and how. And then pick the um, like medium that works for you, uh, whether it's like self led or a course. And third one is like still uh, maintain your knowledge, uh, your expertise and like keep it on a sharp level so that whenever you ask a question you don't feel like you have to go and um, you know, AI it uh, so uh, and then you can still answer the question and have this structure in your head. And for everything that is not important, it's like okay, uh, to uh, just rely on AI. At least for me.

Speaker A: I know because it's beginning so much better. There's so much work that for what a great note to close this out on. Thanks so much Victoria. We have time for one last section and that is plugs. Are there any um, societies or you know, causes near and dear to your heart you want to give a shout out to for our audience?

Speaker B: I think that um, I love whales and dolphins. Uh, so uh, it's just I always uh, tell people to see them, um, and to swim with them. It's becoming more and more popular so just uh, to be ethical about it. Also to never go TO uh, like SeaWorld, uh, because whales and dolphins like need to live in the oceans, uh, but not uh, be uh, in captivity. So and um, like the less people will go to the parks like that the better chances that they will uh, that whales and dolphins will be fully free.

Speaker A: So near and dear to my heart as well. Thank you for that. Awesome. Uh, and from my side, I would like to plug the show so if you enjoy hearing stories from real practitioners who are working in the AI space today, please like and star show whatever, uh, star or like is whatever on your favorite podcast platform. Platform. We will be having our ah, path, uh, to AI Product Leadership, the Indispensable Product Leader masterclasses resuming shortly. So we'll be posting that link too, if you want to come to our free monthly masterclasses where we uncover the path to AI product leadership in this age of AI how to get there, no technical background required, um, to really still leverage and use your expertise in this day and age. So thanks again so much for joining us. Victoria. That's our show for this week. Um, that's another episode of AI Product Leader and we'll see you in another weeks. Couple of couple weeks. Thanks again.

Speaker B: Bye. Thank you, Polly and all the listeners.