Home Blog GenAI Olivier Halupczok: Empowering Web Developers to Create AI-Powered Apps

Olivier Halupczok: Empowering Web Developers to Create AI-Powered Apps

The idea of creating a functional app with AI capabilities in just 30 minutes sounds like a dream, but is it truly unrealistic? And how can Generative AI be effectively used in real-world digital products? These are some of the questions we explore in this week’s “Around the Product Development” webinar, where we sat down with Olivier Halupczok, a GenAI developer at Boldare. Our topic is how to kickstart your first AI-based product. Enjoy!

 Olivier Halupczok: Empowering Web Developers to Create AI-Powered Apps

Table of contents

Matt: Hi everyone, and welcome to Around the Product Development, our weekly 25-minute show on hot topics in digital product creation. We cover everything from monetization to innovation with actionable insights from the Agile Product Builders community. Speaking of which, APBC is a Slack community powered by Boldare, made up of digital product creators and consultants. This week, we have a special guest, Olivier Halupczok, here to talk about how to kickstart your first AI-based product in just 30 minutes. I know that sounds brief, but we’ll cover it all. Olivier is an AI engineer with a background in traditional engineering. Fun fact: Olivier is Polish with a French first name, and he’s lived in France, but these two facts aren’t related, as I recently learned. Welcome, Olivier!

Olivier: Hello, everyone. Thank you for the short introduction. Sure.

Matt: Great to have you here. Thank you so much for joining us today. As I mentioned before, you’re an AI engineer, right? I think it’s something that not all kids dream of, but maybe some do. And perhaps some of our audience would like to know: how do you become an AI engineer? Could you share a bit about yourself-where you come from, your studies, and your professional background - so we can understand the steps one needs to take to become an AI engineer?

Olivier: Yeah, so as you said, I started as a software engineer, and I’ve been interested in software development for a long time. My first job was as a front-end developer, so I’ve developed many applications. This was before I started my studies, so I didn’t have any background related to AI yet. But I was really looking into studying AI and considered applying for programs in the Netherlands and so on. Eventually, I started my studies in Poland, where I’m from, specifically the Silesia region. Now, I’m living in Warsaw and studying IT, specifically computer science. I’m currently in the first year of my computer science studies, and that’s my background.

Matt: Okay, very cool. One last question: as I mentioned before, you have a French name, but you’re not French, right? You have lived in France, though, so I’m curious about that. Where did you live, and why were you there? Was it for studies?

Olivier: No, it was for family reasons, I think, and yeah, it was kind of a surprise for me as well. I have a French name because my parents left France when I was born, and then somehow I ended up in France 13 years later. So that’s my story about France.

Matt: Yeah, maybe it was meant to be, right? With a name like that, you had to end up in France somehow. It was probably how it was supposed to be.

Olivier: Yeah, I think so, and living in Champagne was an excellent experience for me.

Matt: Hey, so let’s get back to the topic. We’re talking about AI solutions, right, and AI-based products and how to kickstart them in 30 minutes. So I think there’s quite a lot to pack into our talk today. To get started, I know Boldare has an AI system that enables very quick building of AI products. Could you tell us a bit about it? What is it, and how does it work? Is it really possible to build a product in 30 minutes? We’ll get to that later, but for someone not familiar with the technology, what does this AI assistant actually do? How does it work?

Olivier: Sure, there are countless ways to incorporate your own AI assistant, and we have developed the technology for that, which I’ll mention later. But there are many ways to integrate AI into various fields of business and software development.

I’ve even thought about whether there is any business field that cannot incorporate AI, and I concluded that in every field of business, there is something that recurs regularly, so there is always a chance to automate it. I think there is always a way to incorporate AI into a business. You can create an AI assistant that helps your users in e-commerce, finance technology, and so on. At Boldare, we have prepared a special open-source library that can quickly kickstart AI development for your product.

Matt: So, as someone who doesn’t really know much about this, if I had to build an AI assistant from scratch, it would be a lot of work. But if I use your AI assistant to build something based on it, it saves a lot of time and effort. Is that correct?

Olivier: Yes, that’s correct. Especially if you have no experience in AI, we’ve prepared a manual that guides you step-by-step through the whole process of bootstrapping and quickly starting your AI assistant. We’ve also created a template that you just need to clone, and then you can start your first AI assistant without any prior background in AI.

Matt: So, talking about the Boldare AI Assistant, right? I understand there’s an AI assistant that helps me save time, and I guess money as well, since time is money. What can I actually do with it, or what can I achieve using it? For me, it’s hard to envision what kind of products I can build with it. Are we talking about very simple things, or are we talking about complex things? What’s even possible with it?

Olivier: You’ve mentioned complex and simple things, so I think I’ll start with the simplest. For example, it can be used for customer service.

Every business has customer service operators, and I think this is one of the best areas to automate. Of course, you can hire customer service agents to work 24 hours a day, but that’s quite expensive. Instead, you can automate it with AI assistance.

You can also add additional features to your product using AI assistance. This technology can automate offer preparation for your customers and provide data automatically, without needing further assistance from any of your employees, for example.

Matt: And then, referring to the second part of our title, we say it’s possible in 30 minutes. But my question is—honestly, it sounds very interesting, right?—to create something in 30 minutes, but how far can one really get in that time? Are we talking about a very simple solution, or is it actually possible to create something more complex in just 30 minutes?

Olivier: It’s great to mention that we already tested this with a developer who wasn’t experienced in AI last week. It took him just 30 minutes to run his first AI agent, or AI assistant. Of course, it was something simple, but it bypassed the need for all the knowledge and familiarity with various tools that would typically be required to build an AI assistant without our open-source AI builder.

He managed to deploy and release his own assistant in 30 minutes for personal purposes, and we’re quite proud of that. After you release your assistant in 30 minutes, you can continue developing it further, add more functions, push it to the limits, and so on. But 30 minutes is enough to create the first version of an AI assistant for your product.

Matt: So, 30 minutes to get started, basically, to have something at least that you can develop further, but not 30 minutes for a fully polished, working product yet, right? It’s more of a very basic MVP, I think.

Olivier: Yes, of course. If you are experienced with this library, you can do much better. We even set a record where we went from idea to a full MVP - or rather, a proof of concept (POC), as we call it—in just 17 minutes. That’s our record for showing an AI custom-tailored for our customer in just 17 minutes.

I think this really shows how much you can achieve with this piece of technology.

Matt: So in 30 minutes, you could have built two products almost, right? At least proof of concepts.

Olivier: Yeah, exactly. Super fast.

Matt: And then maybe my follow-up question, right? Since we’re talking about a developer with experience, you still need a developer, correct? It’s not possible for someone without any development skills to use the Boldare AI Assistant to build something, right? So you need at least a developer. And I think it’s preferable to have an AI developer if you want to do it a bit better and faster, or how does that work?

Olivier: Yeah, it’s really important to know that you need to take advantage of a JavaScript developer because our technology is written in JavaScript. Or we can even take it a step further and call it TypeScript, as this is a subset of JavaScript. So, we wrote it as a JavaScript library.

You might be also interested in the article:

Join Our Discord Server: A Community Hub for our OpenAI Assistant

Join Our Discord Server: A Community Hub for our OpenAI Assistant

And why is that? Because we wanted every web developer—since JavaScript is, of course, a web development language—to be able to create their own AI assistant. Many companies already have JavaScript developers, so we wanted to make this technology accessible to everyone, and that’s how you can incorporate your AI assistant.

I would like to mention that you don’t specifically need a front-end developer or a back-end developer. You just need a JavaScript developer, as both front-end and back-end capabilities can be provided out of the box with our library.

Matt: That’s what I wanted to ask as a second question, right? So any developer who knows JavaScript will be fine in this case, correct? There’s no special expertise needed. Anyone with JavaScript knowledge will be able to handle it.

Olivier: Yes, exactly. Our main goal was basically to take all of the concerns away from the developer and just let them focus on what the AI has to do, what its tasks are, and what tools are needed. That’s everything a developer needs to define to run their first AI assistant, and it just works.

Matt: Right, and then a question I think many people have about AI in general, especially when there’s an AI assistant that helps you build products, is how safe is it actually? How secure is it? Because I assume you might use this AI assistant, but there are many others out there, right? So it’s not just about this one. Let’s make the question a bit broader. You might use them for company-related tasks, which can involve private information. Sometimes user data might be involved, I imagine. You mentioned earlier, for example, if it’s used as a customer support agent, it interacts with customers. Customers might provide data, like email addresses, date of birth, and so on. Is it safe to use, and how secure is it?

Olivier: Oh, that’s a great question. Actually, it’s quite safe. I think it’s very safe because it takes advantage of OpenAI services, which are well-known and trusted nowadays. It’s safe because OpenAI is a reputable enterprise that prioritizes the privacy and security of the data provided by companies from all over the world.

So, we are not concerned about data safety, but there is an even safer option for those who need it. OpenAI’s technology also includes Microsoft Azure services, which are even better for enterprises that are most concerned about data security, customer data, and so on. It’s easy to switch from OpenAI to Azure using the same package, through our package.

Matt: You mentioned OpenAI, right? The company behind ChatGPT that people will know. Well, I think if people are familiar with AI, they might know ChatGPT, they might know OpenAI, and Sam Altman behind it. How is the Boldare Assistant different from just using ChatGPT? Are there any benefits compared to other AI assistants, or how does that work?

Olivier: Basically, our goal was to simplify all the infrastructure and basic tasks you have to handle when using an OpenAI library. This is powered by OpenAI and ChatGPT, of course, but if you want to create your own project using OpenAI, you have to take dozens of steps to accomplish even basic tasks, which can be challenging. I think this can be a friction point for many users of OpenAI libraries and technology.

So, we wanted to make it easier for anyone who wants to have their own OpenAI assistant, and to show them how the project should look from the beginning to make it better. That’s what our library does.

Matt: Let’s say I’m a business owner. I have a business, I have products, right? Let’s say I’m in the e-commerce business—I think that’s a good example. There are lots of ways to use it. What are the next steps I should actually take? I probably need to find a developer. Is there something I can do myself before then, or is the first step to find a developer and move on from there?

Olivier: Sure, one thing the developer needs to have is a clear idea of what the AI has to do. For example, in e-commerce, it could be for customer service. You have to define the tone of the AI—how should our AI Assistant sound? What should its attitude be toward the end user? What tools should it use? Because our AI assistant can use many tools, like fetching a list of products from a database, or taking advantage of user analytics, and so on. It can be super-powered with these tools, but we need a plan for this because our JavaScript developer has to define that in the code.

Matt: I’m not sure if you’re familiar—well, I think you are, right? Could you quickly share a bit about that? I think it’s a good example because it’s still difficult for people to understand what they can do with AI. Maybe you could briefly share how it can help. For example, in this case, it’s an e-commerce business, but maybe you could explain what it actually does. I imagine it probably didn’t work in just half an hour, but even if it takes a few hours to build something, I think it’s an interesting case.

Olivier: Yeah, that was a really interesting case because processing 60,000 to 70,000 products only cost around $30 or $40 per month. So that was a pretty amazing result, I think. What we did using AI was show the user the value of each product on the e-commerce website or platform. We wanted to help users differentiate between two products by showing the advantages and disadvantages of each one. So, we generated three pros and one con for every product across 60,000 products. First of all, it was super fast, and secondly, it was really cheap, costing only about $30 or $40.

Matt: Compared to hiring a writer, right? Because if you have to write 70,000 product descriptions manually, that would be the case, I think, right?

Olivier: Yeah, if you look for a freelancer to do that, it would cost around $150,000.

You might be also interested in the article:

Case study: Expanding Traffic Family's Online Business With AI Support

Case study: Expanding Traffic Family's Online Business With AI Support

Matt: Yeah, because it’s not just about the money and time, but also the opportunity costs, right? You lose time, so you don’t have a product ready in the meantime.

Olivier: Sure, yeah, that’s right. And we can even do it continuously. Every time you add a new product, you can use the description generation as well, so that’s another advantage.

Matt: Okay, I think we’re almost out of time. So I want to ask you two more quick things. How can people get in touch with you or learn more about this AI assistant? What do they need to do? And if people want to know more about the subject, where do you find your information? Where can they find more information?

Olivier: Sure, let me answer the first question. You can always reach out to me through my LinkedIn profile or by email, but not both. For the second question, you can gather more information about Boldare OpenAI Assistant through our blog or on npm.js, which, if you’re a JavaScript developer, you probably know. If you’re not, npm.js is a library provider for the whole language. You can also check out our AI package on GitHub. There’s a complete Readme file and manual for our library there.

And don’t hesitate to reach out if you have further questions. We are eager to help you if you have any concerns about our assistant.

Matt: Perfect, amazing. We’ll share your contact details with the audience later, so they’ll have them, or they can find more information in the resources as well. Olivier, thank you so much for this lovely talk. I don’t know how you feel, but I feel like I need to have an AI assistant at some point or start building something. While I’m quite interested, I still feel like I’m late to the game, and I think many people feel this way. So maybe this is a good time to get started!

Olivier: Yeah, I think this might be surprising, but you can still be ahead of many people who started with AI products recently. This is a really disruptive technology, and it’s evolving so quickly that you can still get ahead of them.

And you can take big steps towards outperforming your rivals, I think, your competitors. Yeah, exactly.

Matt: All right, thank you so much, Olivier, once again. It was amazing having you here.

Join APBC’s product development discussion! Every Monday at 3:00 PM, engage in 25 minutes of conversations with experts.