Introduction and Background
Ryan Davies: Welcome everyone to the Tech Business Roundtable podcast show. This is a podcast show dedicated to shining a spotlight on tech innovators, entrepreneurs, founders, and the compelling narratives behind the movements that they’ve established. I’m your host, Ryan Davies, and I’m hosting today’s discussion on Navigating Tech Startup Waters: A CEO’s Guide to Early-Stage Success with Olga Topchaya Olga. Thank you so much for being here today.
Olga: Thank you for having me.
Ryan Davies: I think this is going to be amazing. I mean, your background is incredible for this topic, and it’s perfect for our audience and our listeners. Olga is the founder and CEO at Lapis AI Consults and specializes in aging in aiding early-stage start-ups and SMBs to embrace AI and for innovation and growth after a vibrant decade-long journey and diver diverse marketing roles. She’s transitioned into the AI domain with a career focused on collaborating with start-ups and SMBs. Olga possesses an intimate understanding of their unique needs, and this has inspired her to establish Lapis. She currently enjoys the equally rewarding role of being a new first-time parent and running a business. It’s really running two very different things there, isn’t it? But somehow, at the same time, both require all of your time, attention, and energy. So I think that’s incredible, and I’d love to just kind of dive right in here a little bit to talk about sort of AI innovation and how you’re really able to help different organizations and different people guide that through those early stages and into the start of success. What I’d love to know, you know, is that it helped cover up that background a bit more for me, and it was what led you to find Lapis Consults and inspired you to focus on the kind of assisting at this level.
Olga: Sounds good. So, how I got here? I had to go on friends leave way earlier than I thought, and then by the time I got back, I was coming back, just like the worst tech recession since the early 2000s. So, yeah, so I decided, ok, instead of me just sending out a bunch of resumes and hearing nothing, it’s not going to happen, right? As we all know, hundreds and hundreds of thousands of people are being like, oh, like, all right, let me take a beat. Here, I hang out with my baby, and it was at that time that I discovered AI Generative AI specifically. It was before the ChatGPT boom. At that point, I think it was GPT three, the early version of GPT three, and I kind of fell in love. I was like, this is amazing. What is this technology? How does nobody know about this? What businesses aren’t using it? Why? And so I just kind of delve into really intense self-study. Anything I can learn about the technology, talking to people, joining Discord groups, you know, my husband’s an engineer, so I would talk his ear off to find out what this is. And that led to my gig as a prompt engineer. It was a short-term contract for an AI company followed by that. That’s when ChatGPT came out, and it just exploded the world, obviously. I led another contract at an AI company, but this time, I focused more on the product and go-to-market aspect. So, applying my marketing background and my understanding of the product and from there, I ended up getting another gig and that kind of like, all right, well, now, you know, now it’s time for an official announcement, and I paired up with a team of just like really phenomenal AI engineers. Over the course of my self-study, I developed quite a few connections with engineers. So, it was already the stage that had been set to fill in that technical gap that I was missing. So I had the business aspect of the product, you know, go to market and marketing, and they are able to assist with actually building the product. So, how do we help businesses? It’s full-service. As I mentioned, we can build a product for businesses that are interested in leveraging their own domain expertise to roll out something new with AI. I can help just integrate AI into their operations into their processes and understand what aspects of their job or their company can be automated. Just training as well. People really don’t know that much about AI, even the ones that do know about AI. For small businesses and start-ups that really want to take their product to market, that’s where I can help them as well.
Navigating Challenges for Early-Stage Start-ups
Ryan Davies: I love that. I think it’s a perfect kind of introduction to what you do, and I’d love to dig a little bit deeper right at that point with early-stage start-ups. Let’s talk. I mean, the challenges are kind of countless, right? There are so many things that they’re doing, trying to balance everything, right? And I think that’s, again, what you can bring to the table. I’d love to get your impression on some of maybe the common hurdles that you’ve observed for these early-stage start-ups and, because you specialize in this, how to integrate AI from precede through like kind of all the way up how AI could really kind of help address some of these challenges.
Olga: So, the absolute biggest mistake that I see early-stage start-ups making is creating something that they think is cool rather than crucial. For it’s like a lot of them, you know, they, especially in AI, you see, they have what they believe is a phenomenal product idea, which it very well might be. But there’s no necessarily market data to back that up. So it’s kind of the reverse of how things should be working when you start a company. first, you identify a market gap or a market need, and then you develop the product to fill that need. A lot of start-ups, I see, just kind of going about it the other way they develop a product, and then they try to figure out who would want that product. For some companies, it works. But, for most, it doesn’t. There are a couple of exercises that you can run yourself, and there are very quick tips I can give you. So, for example, you can run an immediate relief test. What that means is if your product was aiming to solve some kind of a problem, what immediate relief would the customer feel? Right? If it’s automated customer support, will it reduce their customer service wait times? That’s an immediate relief. You can look at the day in a life scenario. How are people going to be using this product on a daily basis? If you’re building a health monitor, you don’t just list out its features, but what solution and problem is it solving? And there’s a bunch of these little tests you can kind of run through to make sure that it, the product isn’t just a cool product, but it is actually going to stick around because it’s something that people need, it’s going to help them solve some kind of a problem. And how AI can come into play is, yeah, there are a lot of ways, but market research is a big one. Now, you can connect your own data. You can connect various APIs to start making those analyses, right? So you’re not just relying on the large language model itself. You can actually connect it and start generating real insights. What are people saying? What do they need? What are the issues that they are being faced with? So, market research is a really big one. Even just from the get-go, you can create very customized personal messaging to reach out to early users. And if you know, you’re really early; you’re looking to get funding, then some of the things that investors are going to look at, they’re going to look at the team who is, you know, who are the people that are actually running the company, creating the company. But they’re really going to be looking at usage and engagement and, you know, recurring engagement. So you want to pick people carefully who are really going to fit that need. So you’re not just wasting your time trying to go off to people who are not going to do anything for you. If so, you would probably start with an early alpha launch. I wouldn’t even call it a beta launch because this is the part where you’re really trying to hone in on your MVP and focusing on those core offerings instead of trying to include this feature and that feature, and I see, you know, companies get stuck in this like endless feature phase as opposed to go and ship, right? Like build a thing, ship it, and then iterate, and that is going to get you much faster to a solid product because then you’re going to have the customer feedback. So, really distinguishing the MVP from a full-on product platform, and this is all for tech, I’m not really talking as much about small businesses; they have a different operation and business model and everything. So we’re all at least sticking to that side of my clients and companies I’m dealing with. But depending on your product, you might have a very, very high ticket item, in which case, you may need less users to show engagement, or you might have a lower ticket item, and so, in that case, you’re going to need to bring on more users but to get those early users, you can start with just within your own team, within your own network, create customized, you can use AI to create those customized messages. You can use it to help you create your ideal customer personas after that. It’s become really, really great at answering questions and offering advice. I didn’t realize that it would get there so fast. But in many ways, you can definitely do it, especially if you’re solid in prompt engineering, and that’s something that we also help with. You can program it right to behave in a certain way. You can give it some baseline information to help you develop those customer personas, navigate those, and then just automate things. You’re doing a million things. I’ll give you an example: when I was starting my company, I would say part of the reason I was able to get up and running so quickly is because I went and created a website using AI and it was fast. It wasn’t perfect. It was far from perfect, right? But it was done, and it was done within 15 minutes. From there, I was like, OK, add this image, and add that image and this page, and it’s better. But I had something. I had a face, quote-unquote, right? Like I wasn’t just Olga Topchaya, you know, get the right. It was Lapis Consult, and we had a team. We have a product description that we could send potential clients to, not just my LinkedIn profile. So these are some of the kinds of things when you’re really in the early stages. You’ll probably have more resources as you get farther along, but for a lot of things, you know, you can also have fewer resources now. So you don’t necessarily need to have a team of 10 engineers. You now have Copilot, so they can do their work faster, right? You can hire, and maybe some of them can even be less experienced, depending on your needs. You know, in our case, I actually need very highly experienced engineers because AI is a new field. They all specialize in AI. So my case is a little bit different, but in some other companies, maybe you don’t need ten like mid-level or entry-level engineers. You can hire a couple of solid engineers, and they’ll do their work for ten engineers using Copilot. So, and these are just some examples, I can probably talk for hours and hours about this.
Ryan Davies: So I was going to say, I mean, like, there’s so many that we could continue to kind of dive through. And I was basically saying, like, you know, are there specific areas where AI tends to give the most impact for start-ups? But I think you’ve talked a lot about that. Maybe we could shift to how you see optimally how start-ups can strategically adopt AI in their operations and foster growth and product development in all of these areas. As you said, I guess it depends on the stage where you’re at, though, right? Like if you appreciate it’s a very different story than somebody who’s in a series or further in terms of of that conversation.
Olga: It really needs to start with an assessment. What is your baseline? What are your goals? Where can you improve? What part of your job do you hate that you wish a magic genie could just take and do for you? That’s a great place to start now. It has to be virtual. I can’t make your ice cream. What part of your job do you do? You hate that maybe it’s just super tedious. And that’s a question that I often will ask my clients who don’t really know what they want to do, but, you know, they want to do something. So it has to start just with a basic assessment as part of things, you know, that we do, which is, you know, part of the reason we don’t like go and build out our own product and we instead help businesses build their products because every business is going to be so different.
Ryan Davies: So you can’t help me with going to get fresh coffee in between recording or anything else, right? But yeah, the rest of it. There’s a long list of things where you’d say, boy, this could help, it saves time, it saves costs like soft costs, hard costs, all of that. So I know that’s a big thing. Right, start-ups have limited resources. You’re trying to build in-house capabilities. But if you’re trying to even build in AI capabilities, that can be daunting, right? So, do you recommend again? Like I know, this is exactly what Lapis does to help navigate the process to make the most of AI external solutions and things like that for somebody who’s even wondering where they can get started. Like you said, that baseline, what does that look like? Is there, is there a checklist that you typically use something along that again? Don’t, don’t give away your secret recipe. That’s what you do, right? But you know, is there sort of that this is how you recognize where you can go?
Olga: So I like to start with understanding what their familiarity is with AI you can do kind of a self-assessment on how advanced you feel you are in prompt engineering, right? Do you know what chain of thought reasoning is? Do you know what meta-prompting is? Plenty of people do, and many people don’t, right? For many people, they don’t even know some of the basic things that AI can do. So they will use it only for content creation, so start really looking within yourself and asking what you know. And what maybe are you missing? Then you can go on to ChatGPT or Anthropic, which is getting much better now, start a conversation, and see, you know, what it thinks, which may be areas in which you can improve your own self-learning. From there, you can look into AI training courses. I would be very cautious with anyone who’s going to say you’re going to make $10,000 in 10 minutes, and here are ten lessons to do it. So stay away from like those youtubers. There’s going to be a lot of those, and the way that you can figure out if a person is really going to offer you quality training or not is they’re going to show you both the not just the prompt, right? So they’re going to show you the input and the output. So right now, what you’re going to see is the 100 best prompts. I don’t know who came up with these things, but they look very reasonable and are great ideas. But what is the input? What is the output? So those are ways that you can get started and even just begin to explore. Ask ChatGPT what you can do, right? What are you capable of? Recently, OpenAI made some new announcements, right? So, see what these big dogs are announcing and what their capabilities are. Amazon, Open AI, Anthropic, and Google Bard all very recently released very advanced capabilities that, a month ago or so, were just not only would I not think that we would be able to do it, but like there were whole entire companies based on creating these kinds of services, right. So, those are some ways that you can get started.
Rayan Davies: I think, given the diversity of start-ups and different business models and things like that, do you recommend customized AI strategies for different stages of growth? Like I’m assuming again, from your expertise, it’s not a one-size-fits-all-all.
Olga: You can use, like, you know, other companies’ experiences for inspiration, which is always great, but it has to be a customized strategy. There are almost no even best practices because it’s so new, right? If you know, we go back to like marketing, which I was doing for a while, at least that also is going to be obviously customized to every company, we will have like some best practices, right? There are no really sustainable best practices, right? What’s the best practice today does not apply 2-3 weeks from now, so it has to be very fluid. You have to accept the fact that whatever it is you’re going to do now is going to be different in a month. If you want to play in this game, you need to be flexible, and that, I think, is where early-stage start-ups and small businesses are actually really uniquely positioned because they can just change on a moment’s notice, they’re not, you know, they don’t have responsibilities to all these shareholders and being I mean, it’s just obviously you have like 100 people that need to make a decision before a single email can get sent out. So I think in that way startups are really going to soar if they can embrace that flexibility.
Ethical Considerations and Responsible AI Use
Rayan Davies: Yeah, I think it’s always the hardest thing, right, is change management and that, like you said, when it’s a start-up, the smaller you are, the easier it is to change something and pivot on a whim or on a dime or because you see something pretty hard to say we’re implementing this. So we’re going to go, and this is going to affect 1500 people in a company, and we’ll hope it works if it doesn’t. Well, you know, you can’t really backtrack and get the address and everything else. So that’s just incredibly crucial from that perspective. I think one of the big things again for our tech business audience here, you know, a lot of them, again, they’re involved in tech, they understand it. This is their world, but there’s still this kind of fear-mongering going on around the ethical considerations that AI brings and all of that, and I think that’s probably part of the reason why some people might be hesitant to bring it in, but like that’s crucial. How do you guide start-ups to ensure this responsible and ethical use of AI technologies in their products and services and what they’re looking to do?
Olga: Yeah, absolutely. I’m really glad that you brought that up. I will start with saying that there is now new legislation that has given some guidance, and Europe passed their own act, and the Biden administration, the US here passed the executive order. You can feed it into ChatGPT and have it give you a summary. So it’s a start, right? But it’s only a start. A lot of it, when you read it, it is like, OK, we plan to have a plan kind of thing. So it’s a start, and you know, for ethical purposes, it’s essentially anything that’s ethical, do not harm. That’s how you can summarize all the ethical do not harm. So you can take certain precautions to eliminate biases if you are creating your own product. So, there’s a methodology called Retrieval Augmented Generation or RAG. In this case, you can feed your own documents into it. You can say, OK, respond like this is what I want you to say. This is what I want you to do using only this knowledge base, right? And if you don’t know, so you don’t know. So that is now a methodology that is available to us that was not necessarily available beforehand, right? Being transparent with users on how you are using their data, right? If you are collecting any kind of data. I think companies are already more accustomed to it. You know, over the last few years, we have had GDPR and GDPA, right? We have all those laws already in place. So keep thinking about that, right? It’s the same kind of mindset. To be completely blunt, will AI take some jobs? Yes, it will. It’s more likely going to take jobs from people who are not very good at their jobs but the people who are good at their jobs who can make use of the technology. They’re going to be just so much better because they can do so much more. If you want your employees to know how to use the technology, then you can make them better, you can guide self-development, and you can guide better decision making. That’s great. Right? Like that’s great for business. You are just being able to do more and then in terms of the fear aspect, which is completely understandable. Like it’s this technology, it was just almost sprung upon us, and we had no war in a way, right? Like, except for the people that were in AI and so and it just got so quick and just like exponential evolution, it’s completely understandable that people are afraid, but one way that I urge people to think about and actually just co-authored an article on this, like yesterday was published is a parent-child framework. Now, I’m not to say AI is a sentient or anything like that. No, that’s not the case.
Ryan Davies: It’s a whole series of events.
Olga: Yeah, we’re not, we’re not doing that. It’s only for a practical way of thinking about technology. It’s not your nemesis, right? It’s a thing that we created, and we need to be responsible for quote-unquote raising, right? It’s been in the process for a long time, the catalyst, right? This boom happened only recently, but AI and ML have been around for a very long time, right? So, kind of keeping in mind that it’s an evolution that was very deliberate, and we have to be responsible for using it, and developers have to be responsible for developing it, right? You know, just like a parent, right? You are responsible for the things that you make, and if you’re not acting responsibly, then CPS is going to come after you.
Ryan Davies: and it’s nearly impossible to undo something where you’ve made, you know, push it too far.
Olga: Right. I think adapting this framework where you’re viewing it as something that you’ve made can help you, right? That’s a partner or a child, a progeny, right? But it’s it’s not in nemesis, right? I think maybe we will help eliminate some of the fear and yes, keep in mind it’s not sentient, it’s not by any means, it doesn’t have the motions or anything, right? Can simulate things, but we are not a point of AGI. We probably will be, but we’re not there right now with the technology as it stands, right?
The Future of AI in Start-ups and Preparation Strategies
Ryan Davies: I think that leads into a great kind of outro question here as we’re kind of shifting ttowardthe end of the episode. What role do you foresee AI playing in the future of start-ups? And really, how can early-stage companies best prepare to continue to integrate AI into the business landscape?
Olga: So I will refrain completely from making any kind of predictions in this realm. It is impossible. Even the big dogs don’t know, and we’re probably fairly close to AGI that much, I can say, but beyond that, I don’t know. The best way to prepare is to plan for change. So, and you know, I guess, assume that whatever it is that you’re doing right now can, at a moment’s notice, be something completely different, and when that happens, how are you going to react? You’re going to react with panic or are you going to take a step back and then start thinking? OK, cool. Now, how do I use this to my advantage? I really do urge companies not to adopt the business model where everything that they do is only reliant on AI, which I think is very dangerous because then you don’t control your product. It means you don’t control your business. So, I would say establish your domain, right? If your domain is in construction in tech, right? On the tech side, right? You know, or like automation, if your domain is in health tech, if your domain is fintech, right? That’s where you’re playing and start thinking how can I shift that business side ecom is also like a great example, right? If things are beginning to change, you see where the winds are blowing a little bit, right? You see that things are kind of going on the assistant side, right? So start planning for change in whatever scenario that may be.
Ryan Davies: I think that’s perfect advice to kind of close off here, and I would love to, just at the very end here, get people to connect with you again. I think we’ve got a lot of being a Tech Business Roundtable. We’ve got a ton of founders, we even have like, you know, VCs and information on that, all of that, how to connect with you and really, you know, help navigate those start-up waters that we’ve been talking about to make sure that their early stages turn into the next success story.
Olga: Yes, so going to the website lapisconsults.com or going to LinkedIn is probably my personal LinkedIn. It is going to get you the fastest response, which is Olga Topchaya. You can reach out to me directly or send me an email at olga@lapisconsults.com.
Ryan Davies: Perfect. I think we’re going to have some people taking advantage of that for sure, and it’s just that she’s the only one who is easy to find. I mean, so much good advice there for you to be able to take in, you know, again, find that baseline, start to deploy it, make, make scalability faster, easier with lower cost, everything that we’re looking for that, I think startups are so desperate to be able to to do to keep up and stay ahead. So Olga, thank you so so much for being with us today and sharing your information like we said, think a couple of times ago, we could do a whole podcast on this episode. We might have to have a whole. We will just get you your own podcast. The next strategy is to do that, and we’ll have you do that. So that’s perfect. Thank you so much for being here.
Olga: Thank you, Ryan.
Ryan Davies: Thank you. Excellent. So I want to thank Olga again for this amazing podcast on Navigating Tech Startup Waters: A CEO’s Guide to Early-Stage Success, really took a deep dive into how AI is going to be able to get help, level you up, help you scale up and take away those things you can. Like you said, the genie that you’d love to get rid of, but do all of the things that the start-ups crave to be able to do to get ahead. So, thank you for that, and also want to thank our listeners. We can’t do what we do without you. So, until we meet again with another amazing TBR episode, I’m your host, Ryan Davies. Thanks, everybody. Take care.